J Neurol (2017) 264:856–861

DOI 10.1007/s00415-017-8429-8

ORIGINAL COMMUNICATION

Diagnosis of Guillain–Barré syndrome in children and validation

of the Brighton criteria
Joyce Roodbol1,2 • Marie-Claire Y. de Wit2 • Bianca van den Berg1,5 •
Vivienne Kahlmann1,2 • Judith Drenthen1,4 • Coriene E. Catsman-Berrevoets2 •

Bart C. Jacobs1,3

Received: 4 January 2017 / Revised: 20 February 2017 / Accepted: 20 February 2017 / Published online: 1 March 2017
Ó The Author(s) 2017. This article is published with open access at Springerlink.com

Abstract To describe the key diagnostic features of seen in 97%, including 5 children with a treatment-related
pediatric Guillain–Barré syndrome (GBS) and validate the fluctuation. Two children had a later relapse at 9 weeks and
Brighton criteria. Retrospective cohort study of all children 19 weeks after onset. 77% of the children showed an ele-
(\18 years) diagnosed with GBS between 1987 and 2013 vated protein level in CSF. Nerve conduction studies
at Sophia Children’s Hospital, Erasmus MC, Rotterdam. showed evidence for a poly(radiculo)neuropathy in 91% of
Clinical information was collected and the sensitivity of the the children. 46 children had a complete data set, the
Brighton criteria was calculated. 67 children (35 boys) sensitivity of the Brighton criteria level 1 was 72% (95%
were included, with a median age of 5.0 years [interquar- CI 57–84) and 96% (95% CI 85–99) for level 2 and 98%
tile range (IQR) 3.0–10.0 years]. Bilateral limb weakness (95% CI 88–100) for level 3. The majority of the pediatric
was present at hospital admission in 93% of children, and GBS patients presented in this cohort fulfilled the current
at nadir in all patients. Children presented with tetraparesis diagnostic criteria.
in 70% or with paraparesis in 23%. Reduced reflexes in
paretic limbs were observed at hospital admission in 82% Keywords Guillain–Barré syndrome
Pediatrics

and during follow-up in all children. The progressive phase Brighton criteria
Cerebrospinal fluid
Nerve conduction
lasted median 6 days (IQR 3–8 days) and less than study
4 weeks in all children. A monophasic disease course was

Electronic supplementary material The online version of this

article (doi:10.1007/s00415-017-8429-8) contains supplementary
material, which is available to authorized users.

& Bart C. Jacobs 1

Department of Neurology, Erasmus MC-Sophia Children’s
b.jacobs@erasmusmc.nl Hospital, University Medical Center Rotterdam,
Joyce Roodbol PO box 2040, 3000 CA Rotterdam, The Netherlands
j.roodbol@erasmusmc.nl 2
Paediatric Neurology, Erasmus MC-Sophia Children’s
Marie-Claire Y. de Wit Hospital, University Medical Center Rotterdam, Rotterdam,
m.c.y.dewit@erasmusmc.nl The Netherlands
3
Bianca van den Berg Immunology, Erasmus MC-Sophia Children’s Hospital,
b.vandenberg.2@erasmusmc.nl University Medical Center Rotterdam, PO box 2040,
3000 CA Rotterdam, The Netherlands
Vivienne Kahlmann
4
Vivienne.kahlmann@gmail.com Clinical Neurophysiology, Erasmus MC-Sophia Children’s
Hospital, University Medical Center Rotterdam, Rotterdam,
Judith Drenthen The Netherlands
j.drenthen@erasmusmc.nl
5
Department of Neurology, St. Elisabeth Hospital, Tilburg,
Coriene E. Catsman-Berrevoets The Netherlands
c.catsman@erasmusmc.nl

123
J Neurol (2017) 264:856–861 857

Introduction A clinical fluctuation was previously defined as an

improvement or stabilization longer than 1 week followed
The Guillain–Barré syndrome (GBS) is a clinical diagno- by secondary deterioration of at least one grade in the GBS
sis, supported by the results of the cerebrospinal fluid disability score [8, 9]. A treatment-related fluctuation
(CSF) and nerve conduction studies (NCS) [1]. Recogni- (TRF) was defined previously as a clinical fluctuation due
tion of GBS is important to start treatment and monitoring to the transient effect of the treatment that usually occurs
as soon as possible. Accurate diagnostic criteria for GBS within 8 weeks after start of treatment [8].
are also required to determine background incidence rates CSF was examined for protein level and for cell count.
and to conduct vaccine safety studies. For this purpose, All results of the NCS were reviewed by a clinical neu-
criteria for GBS were developed in 1978 by the National rophysiologist (JD) and were considered supportive of the
Institute of Neurological and Communicative Disorders clinical diagnosis if consistent with the criteria for acute
and Stroke (NINDS) and updated in 1990 [2]. In response inflammatory demyelinating polyneuropathy (AIDP), acute
to the H1N1 vaccination campaign in 2009 and its possible motor axonal neuropathy (AMAN), acute motor axonal
relation with GBS, new case definitions for GBS were sensory neuropathy (AMSAN) or inexcitable nerves [10].
developed by the Brighton Collaboration, an international An equivocal electrophysiological result not meeting the
collaboration sponsored by the World Health Organization
to improve vaccine safety monitoring [3]. Previous studies
Table 1 Neurological deficits at admission and nadir in 67 children
have validated the Brighton criteria in cohorts of adult with GBS
patients with GBS from South Korea and The Netherlands
Demography Presentation
[4, 5]. The Brighton criteria also require validation for GBS
in children, which may differ considerably from GBS in Male/female ratio 35/32
adults [6]. Age at admission (years)a 5 (IQR 3–10, range 0–16)
In 2011 the Brighton criteria were validated in children Antecedent events
from India, but the clinical presentation of GBS in India is Diarrhea 40% (25/64)
not the same as in a Western country. Therefore, in this Upper respiratory tract infection 41% (27/66)
study, we describe in detail the clinical presentation and Vaccination 8% (5/67)
course of GBS in children, focussing on the key diagnostic
Admission Nadir
characteristics, and validate the Brighton criteria for pedi-
atric GBS. Neurological signs and symptoms
Cranial nerve deficits 53% (32/61) 60% (39/65)
Sensory deficits 30% (12/40) 40% (19/48)
Patients and methods Pain 73% (47/64) 84% (54/64)
Bilateral weakness 93% (55/59) 100% (65/65)
Medical files and discharge letters from all children Tetraparesis 70% (40/57) 88% (57/65)
(\18 years) diagnosed with GBS at Sophia Children’s Paraparesis (of the legs) 23% (13/57) 11% (7/65)
Hospital, Erasmus MC, Rotterdam, The Netherlands, Decreased reflexes in weak limbs 82% (45/55) 100% (62/62)
between 1987 and 2013, were reviewed retrospectively. Autonomic dysfunction 12% (8/67) 53% (35/66)
Part of this cohort has been described previously for dif- Blood pressure fluctuations 6% (4/66) 35% (23/65)
ferent purposes [6]. The revised version of the NINDS Cardiac dysrhythmia 3% (2/67) 15% (10/66)
diagnostic criteria from 1990 [2] was used as guideline for Bladder dysfunction 6% (4/65) 25% (16/64)
the diagnosis. Patients with Miller Fisher syndrome, acute GBS disability scoreb
onset chronic inflammatory demyelinating polyneuropathy 2 35% (23/66) 8% (5/67)
(CIDP) or other neurological diseases or comorbidity 3 26% (17/66) 16% (11/67)
influencing the GBS diagnosis were excluded. Data were 4 36% (24/66) 51% (34/67)
collected regarding age, sex, preceding events, onset of
5 3% (2/66) 24% (16/67)
weakness, neurological signs and symptoms at hospital
6 0% (0/66) 2% (1/67)
entry, clinical course, and results from CSF examination
and NCS. Severity of the disease at nadir was defined by 2 able to walk 10 m unaided, unable to run, 3 able to walk 10 m with
the highest GBS disability score (Table 1) [7]. aid, 4 Bedridden or chairbound, 5 requiring assisted ventilation, 6
deceased
Symmetrical weakness was defined as the absence of a
Median (interquartal range and full range)
difference in muscle weakness in major limb muscle b
GBS disability score
groups on the left versus the right side.

123
858 J Neurol (2017) 264:856–861

criteria for one of these subtypes, but still consistent with persistent paraparesis of the legs during follow-up,
the diagnosis GBS, was recorded separately [5]. although three of them developed reduced reflexes of the
The Brighton criteria consist of four levels of diagnostic arms.
certainty. Level one has the highest diagnostic certainty;
these patients fulfil all diagnostic criteria. Level 4 has the Reflexes
lowest diagnostic certainty, these patients do not fulfil the
criteria of level 3 but all other diagnoses are excluded. The At admission, most children had decreased reflexes in weak
diagnostic criteria needed to fulfil each level are shown in limbs (Table 1). One child initially had hyperreflexia in
online only Table 1. All patients in this study were clas- weak limbs with plantar reflexes and a normal MRI of
sified according to the Brighton criteria. This was done for cerebrum and myelum, excluding transverse myelitis. This
the entire cohort and also in a subgroup of patients in patient developed hyporeflexia during the course of the
whom all clinical information regarding the six key diag- disease.
nostic features were present. Sensitivity of the Brighton
criteria was calculated for the levels 1, 2 and 3. Course of the disease
The study was approved by the medical ethical review
committee of the Erasmus MC. All children reached nadir within 28 days of onset of
weakness and 43 (66%) children already within 1 week. 65
Statistical analysis children (97%) had a monophasic disease course. Five
(8%) children had a clinical fluctuation within 8 weeks of
Continuous data were presented as means and standard onset of weakness, which were interpreted as TRF but
deviations if normally distributed, and as medians and considered compatible with a monophasic disease course.
interquartile ranges (IQR) when not normally distributed. Two of these children received a second course of IVIg
Categorical data were presented as proportions. Shapiro– because of the severity of the deterioration. Two (3%)
Wilk test was used to define if the data were normally children had a clinical fluctuation more than 8 weeks after
distributed. Continuous data were compared with t test if onset of weakness. One child deteriorated at day 59 but the
normally distributed and with Mann–Whitney U test if not weakness was milder than the first episode. The other child
normally distributed. Proportions were compared using the had a clinical fluctuation at day 137 during a Shigella
Chi-square or Fisher exact test. Correlations between cat- gastro-enteritis. Both children recovered spontaneously and
egorical data were tested using the Spearman correlation. no further fluctuations occurred.
SPSS Statistics 20.0 was used for the statistical analyses. A One child had a relapsing GBS with two recurrences at 3
two-sided p value \0.05 was considered to be statistically and 23 years after the first episode. She was treated with a
significant. course of IVIg after each episode with a good clinical
response and was stable without residual symptoms
between these episodes. One patient died during the course
Results of the disease due to severe autonomic dysfunction.

The clinical features of 67 children diagnosed with GBS Treatment

included in the study are shown in Table 1. The key
diagnostic characteristics important for the Brighton cri- In this cohort, only 6 children (9%) did not receive specific
teria are described in more detail. treatment for GBS. The majority of the children received
either IVIg alone (N = 49, 77%) or plasma exchange alone
Muscle weakness (N = 1, 2%) or a combination of IVIg and prednisone/
methylprednisolone (N = 8, 13%). This last category of
At admission all children presented with symmetrical limb patients participated in the randomized controlled trial
weakness, except four cases (6%). These four children comparing IVIg and methylprednisolone (MP) versus IVIg
without weakness initially presented with different symp- and placebo (Koningsveld et al. 2004) [11]. Some patients
toms or signs, including neck stiffness, facial weakness and received IVIg and MP shortly after finalizing the trial
limb muscle pain. In the following days, all developed a considering that this combination was related with better
symmetrical tetraparesis and were diagnosed with GBS. At outcome after adjustment for age and GBS disability score.
hospital admission, 13 (23%) children presented with One of these eight children had increased intracranial
weakness restricted to the legs. During the course of the pressure during the acute phase of GBS for which he
disease, six of them developed additional weakness of the received IV steroids.
arms. The remaining seven (11%) children showed

123
J Neurol (2017) 264:856–861 859

CSF examination the results indicated a demyelinating polyneuropathy of both

arms and legs. The electrophysiological subtype was not
In 62 (93%) patients a lumbar puncture was performed, but related to the interval between onset of weakness and per-
only in 57 (85%) patients the results on both cell count and forming the NCS.
protein level were available. The interval between onset of
weakness and lumbar puncture was median 4 days (IQR Brighton classification
2–8 days, range 0–26 days). A mild pleocytosis was
observed in 27 children (47%); the cell count ranged The classification according to the Brighton criteria depends
between 5 and 10 leukocytes/ll in 15 children, 11 and 20 on the completeness of the data regarding the key diagnostic
leukocytes/ll in 10 children and 21 and 50 leukocytes/ll in characteristics. Therefore, the criteria were validated sepa-
3 children. Two children showed an increased cell count rately for the subgroup of patients with a complete dataset and
with 54 and 60 leukocytes/ll, but both children may have for the total group of patients (Online Resource Table 2). In
had a traumatic puncture; no other diagnosis was made. the 46 children with a complete dataset, 33 children reached
CSF protein level was increased in 46 (77%) of the patients level 1, 11 children reached level 2, one child reached level 3
(Fig. 1a). Of the children with a raised CSF protein level, and one child reached level 4. From the children who did not
at least four had a traumatic puncture. Most CSF samples reach level 1, 8 children (17%) either had a normal CSF
obtained more than 1 week after onset of weakness showed protein level or normal NCS, 3 children (7%) had both a
an increased protein level. normal CSF protein level and a normal NCS. The child who
reached level 3 had more than 50 leukocytes in CSF. The
Nerve conduction studies patient who only reached level 4 did not have a monophasic
disease course. The sensitivity was 72% (95% CI 57–84) for
In 53 (79%) patients data from NCS were available for the level 1, 96% (95% CI 85–99) for level 2 and 98% (95% CI
current study (Fig. 1b). The median number of days between 88–100) for level 3. Patients with various Brighton levels did
onset of weakness and NCS was 9 days (IQR 5–14 days, not differ regarding age, sex, preceding event, disease
range 1–37 days). In 48 (91%) children, the NCS showed severity and outcome. As expected, the Brighton criteria had
evidence for the presence of a poly(radiculo)neuropathy and a lower sensitivity in the whole group of patients because of
supported the diagnosis GBS. In the remaining five (9%) missing data (Online Resource Table 2).
children, the NCS were normal and not repeated. AIDP was
the predominant subtype present in 53 children (60%), fol-
lowed by AMAN in 2 children (4%), both AMSAN and Discussion
unresponsive nerves in one child (2%) In 12 children (23%),
the abnormal NCS were equivocal and could not be further In the current study, we described in detail the variation in
classified. Of the 7 children with persistent paraparesis, NCS key clinical characteristics in 67 children with GBS. The
of the arms was performed in 2 children, and in both patients far majority showed a rapidly progressive flaccid

A B

Fig. 1 Diagnostic test results. a Frequency of increased protein levels 1–10 years: 0.16–0.31 g/L, 10–18 years: 0.24–0.49 g/L. %, the
in CSF in children with GBS. Reference values CSF protein in percent of patients with an increased CSF protein level. b Results
children used in Sophia children’s hospital: 1–3 months: 0.24–0.65 g/ nerve conduction studies in children with GBS
L, 3–6 months: 0.23–0.37 g/L, 6–12 months: 0.17–0.35 g/L,

123
860 J Neurol (2017) 264:856–861

tetraparesis with reduced reflexes reaching nadir within limitation of all studies investigating the performance of the
4 weeks, followed by slow recovery without relapses. The Brighton criteria in pediatric GBS, including ours, was the
diagnosis was confirmed in most cases by the presence of retrospective design and the influence of missing data.
an increased protein level in CSF and findings compatible Additional investigations of CSF and NCS were not
with a poly(radiculo)neuropathy in NCS. In a minority of performed in ten children (15%). Most frequently, these
children, the presentation differed from this prototypic studies were not conducted because the patient had a mild
form. First, 13 (23%) children presented with paraparesis form of GBS or no alternative diagnosis was suspected.
of the legs and seven (11%) showed a persistent para- Physicians may also be more reluctant to perform invasive
paresis during the entire course of disease. In none of these or painful investigations in children than adults. Less-in-
patients, there was evidence for myelum involvement vasive techniques than lumbar puncture and NCS may be
based on clinical examination, MRI of the spine (conducted considered to confirm the diagnosis of GBS. Recently,
in 2 of 7 cases), or clinical re-examination during follow- gadolinium-enhanced magnetic resonance imaging of the
up. A similar paraplegic variant of GBS has been reported nerve roots was reported to be equally accurate as NCS
previously in adult patients (18). Second, CSF protein level [13], and lumbar puncture [14]. MRI may be especially
was normal in 14 (23%) children, and a CSF pleocytosis valuable in centers with limited pediatric neurophysiolog-
between 5 and 50 leukocytes/ll was observed in 28 (49%) ical expertise. Other diagnostic techniques potentially
children. Third, results of NCS were equivocal in 12 (23%) useful in children are the compound muscle action poten-
and normal in 5 (9%) children, implying that in one-third tial scan (CMAP scan) [15] and nerve ultrasound [16]. To
the electrophysiological subtype could not be defined in a further improve the diagnostic criteria for GBS, the
single NCS using current criteria. Additional diagnostic specificity needs to be defined in children with similar
work-up and a follow-up of at least 6 months in these clinical presentations as GBS but an alternative diagnosis.
atypical cases revealed no alternative diagnosis, indicating In addition, prospective studies are required including
that these variations are part of the spectrum of phenotypes patients with the full spectrum of subforms of GBS. The
within the diagnosis of GBS. Children with a complete Brighton criteria were developed primarily for surveillance
dataset reached Brighton level 1 in 72%, and at least level 3 and vaccine safety studies rather than for clinical decision-
in 98%, indicating that the Brighton criteria have a high making in the diagnostic work-up in clinical practice in
sensitivity for the diagnosis of GBS in children. These individual patients. Development of protocols for routine
results show that children usually present with the classic diagnostic work-up and criteria for early diagnosis would
symptoms of GBS and fulfill the current diagnostic criteria. support the early diagnosis of GBS in children.
Accuracy of the Brighton criteria developed for vaccine
safety monitoring is especially relevant for children who are Compliance with ethical standards
frequently exposed to vaccinations. Two previous studies Conflicts of interest Dr. J. Roodbol has no disclosures to report. Dr.
have reported on the sensitivity of the Brighton criteria for M.C.Y. de Wit received honoraria paid to her institution by Novartis
GBS in children [4, 12]. A study from India was based on the for serving on a steering committee and presenting at a conference,
national polio surveillance program in children (\15 years) and has received research funding from the Nationaal Epilepsie Fonds
(Dutch epilepsy fund), Hersenstichting and Sophia Foundation. Dr. B.
and selected 79 (11%) patients with a full diagnostic workup van den Berg has no disclosures to report. Dr. V. Kahlmann has no
from an original population of 718 children diagnosed with disclosures to report. Dr. J. Drenthen has no disclosures to report. Dr.
GBS [12]. This study showed a comparably high frequency of C.E. Catsman-Berrevoets has no disclosures to report. Dr. B.C. Jacobs
patients reaching level 1 (62%) or at least level 3 (86%), and a has received research funding from the Netherlands Organization for
Health Research and Development, Erasmus MC, Prinses Beatrix
similar frequency of patients with normal CSF results (16%) Spierfonds, Stichting Spieren voor Spieren, GBS-CIDP Foundation
and equivocal NCS (29%). The authors indicated that the International, CSL-Behring and Grifols.
investigated population in their study was likely biased
towards more severe cases who more frequently get a full Ethical standard The authors declare that the research documented
in the submitted manuscript has been performed in accordance with
diagnostic work-up. In a study from South-Korea, none of 18
the ethical standards (Declaration of Helsinki 1964) and has been
children reached level 1, but all reached level 2 or 3 [4]. approved by the appropriate ethics committee of their hospital.
Compared to a previous study in adult patients from The
Netherlands, the Brighton criteria are more sensitive for the Open Access This article is distributed under the terms of the
diagnosis of pediatric GBS [5]. Only 61% of adult patients Creative Commons Attribution 4.0 International License (http://crea
tivecommons.org/licenses/by/4.0/), which permits unrestricted use,
reached level 1, compared to 72% of children in the current
distribution, and reproduction in any medium, provided you give
study. This difference is almost fully explained by the lower appropriate credit to the original author(s) and the source, provide a
frequency of an increased protein level in CSF in adults link to the Creative Commons license, and indicate if changes were
compared to children, despite the fact that the timing of the made.
lumbar puncture in adults and children was similar. A

123
J Neurol (2017) 264:856–861 861

References 9. Ruts L, van Koningsveld R, van Doorn PA (2005) Distinguishing

acute-onset CIDP from Guillain-Barre syndrome with treatment
1. van den Berg B, Walgaard C, Drenthen J, Fokke C, Jacobs BC, related fluctuations. Neurology 65:138–140
van Doorn PA (2014) Guillain-Barre syndrome: pathogenesis, 10. Hadden RD, Cornblath DR, Hughes RA, Zielasek J, Hartung HP,
Toyka KV, Swan AV (1998) Electrophysiological classification
diagnosis, treatment and prognosis. Nat Rev Neurol 10:469–482
2. Asbury AK, Cornblath DR (1990) Assessment of current diag- of Guillain-Barre syndrome: clinical associations and outcome.
nostic criteria for Guillain-Barre syndrome. Ann Neurol Plasma exchange/sandoglobulin guillain-barre syndrome trial
27(Suppl):S21–24 group. Ann Neurol 44:780–788
11. van Koningsveld R, Schmitz PI, Meche FG, Visser LH, Meulstee
3. Sejvar JJ, Kohl KS, Gidudu J, Amato A, Bakshi N, Baxter R,
Burwen DR, Cornblath DR, Cleerbout J, Edwards KM, Heininger J, van Doorn PA, Dutch GBS study group (2004) Effect of
U, Hughes R, Khuri-Bulos N, Korinthenberg R, Law BJ, Munro methylprednisolone when added to standard treatment with
U, Maltezou HC, Nell P, Oleske J, Sparks R, Velentgas P, Ver- intravenous immunoglobulin for Guillain-Barre syndrome: ran-
meer P, Wiznitzer M, Brighton Collaboration GBSWG (2011) domised trial. Lancet 363:192–196
Guillain-Barre syndrome and Fisher syndrome: case definitions 12. Mateen FJ, Cornblath DR, Jafari H, Shinohara RT, Khandit D,
and guidelines for collection, analysis, and presentation of Ahuja B, Bahl S, Sutter RW (2011) Guillain-Barre Syndrome in
immunization safety data. Vaccine 29:599–612 India: population-based validation of the Brighton criteria. Vac-
4. Choe YJ, Cho H, Bae GR, Lee JK (2011) Guillain-Barre syn- cine 29:9697–9701
drome following receipt of influenza A (H1N1) 2009 monovalent 13. Smith N, Pereira J, Grattan-Smith P (2014) Investigation of
vaccine in Korea with an emphasis on Brighton Collaboration suspected Guillain–Barre syndrome in childhood: what is the role
for gadolinium enhanced magnetic resonance imaging of the
case definition. Vaccine 29:2066–2070
5. Fokke C, van den Berg B, Drenthen J, Walgaard C, van Doorn spine? J Paediatr Child Health 50:E72–E76
PA, Jacobs BC (2014) Diagnosis of Guillain-Barre syndrome and 14. van den Berg B, Fokke C, Drenthen J, van Doorn PA, Jacobs BC
validation of Brighton criteria. Brain 137:33–43 (2014) Paraparetic Guillain–Barre syndrome. Neurology
6. Roodbol J, de Wit MC, Walgaard C, de Hoog M, Catsman- 82:1984–1989
Berrevoets CE, Jacobs BC (2011) Recognizing Guillain-Barre 15. Drenthen J, Maathuis EM, Ruts L, van Doorn PA, Blok JH,
syndrome in preschool children. Neurology 76:807–810 Visser GH (2008) Serial CMAP scan analysis in Guillain-Barre
7. Hughes RA, Newsom-Davis JM, Perkin GD, Pierce JM (1978) patients. J Peripher Nerv Syst 13:167
Controlled trial prednisolone in acute polyneuropathy. Lancet 16. Goedee HS, Brekelmans GJF, van Asseldonk JTH, Beekman R,
2:750–753 Mess WH, Visser LH (2013) High resolution sonography in the
8. Ruts L, Drenthen J, Jacobs BC, van Doorn PA, Dutch GBSSG evaluation of the peripheral nervous system in polyneuropathy - a
review of the literature. Eur J Neurol 20(10):1342–1351
(2010) Distinguishing acute-onset CIDP from fluctuating Guil-
lain-Barre syndrome: a prospective study. Neurology
74:1680–1686

123