CSF tau protein: A Abstract—We measured the CSF tau protein levels in 26 patients with

Guillain-Barré syndrome. The levels of the poor outcome group (Hughes grade
new prognostic at 6 months was between II and VI, n ⫽ 6) were higher than those of the good
marker for Guillain- outcome group (0 or I, n ⫽ 20) (p ⬍ 0.0005). The higher levels of CSF tau may
reflect axonal degeneration and could predict a poor clinical outcome in
Barré syndrome Guillain-Barré syndrome.
NEUROLOGY 2006;67:1470–1472

K. Jin, MD; A. Takeda, MD, PhD; Y. Shiga, MD, PhD; S. Sato, MD, PhD; A. Ohnuma, MD, PhD;
H. Nomura, MD, PhD; H. Arai, MD, PhD; S. Kusunoki, MD, PhD; M. Ikeda, PhD; and Y. Itoyama, MD, PhD

Since the introduction of plasma exchange or IV im- between II and VI were classified into the poor outcome group
(n ⫽ 6). Antecedent episodes in the 4 weeks before the onset of
munoglobulins as standard therapies for Guillain- symptoms were determined. Electrophysiologic studies were con-
Barré syndrome (GBS), the general prognosis has ducted within 14 days from the onset. Pretreatment serum and
dramatically improved. Although most patients with CSF samples obtained within 24 days from the onset were frozen
GBS treated show good recovery from neurologic im- and stored at – 80 °C for subsequent analyses.
Electrophysiologic studies. Nerve conduction studies were
pairments, 11% of GBS cases still die and 16% have performed using standard procedures with a Neuropack electro-
long-term disability.1 A poor outcome in GBS is cor- myographic machine (Nihon Kohden, Tokyo, Japan). According to
related with some clinical findings, such as older age the electrodiagnostic criteria,8 patients were classified into five
at onset, longer time to nadir, necessity for ventila- categories; normal, demyelinating, axonal, inexcitable, and
equivocal.
tory support, presence of antecedent diarrhea, and Laboratory analyses. The presence of antibodies for glycolipid
electrophysiologic signs of axonal degeneration.1 antigens (GalNAc-GD1a and GM1) was determined by an enzyme
However, it is generally difficult to predict the clini- linked immunosorbent assay (ELISA) because they were reported
to be a good marker for axonal degeneration in GBS.9 The serum
cal outcome in the early phase of GBS. was considered to be seropositive for each of the antigens when
Tau protein (tau) is a microtubule-associated pro- the corrected OD was more than 0.1. The CSF tau levels were
tein primarily localized in neuronal cells. In dam- measured using an ELISA kit, Fino Scholar hTAU (Innogenetics,
aged brain, tau is released into the CSF from the Ghent, Belgium).2
Statistical analyses. The Mann–Whitney U test was used for
neuronal cytoplasm and the CSF tau levels are the comparison of age at onset, time to nadir, Hughes grade, and
known to be a good diagnostic marker for several CSF findings. The Fisher’s exact probability test was used for the
CNS disorders.2-4 Although tau also exists in the comparison of some other clinical features. Both univariate and
multivariate models were developed for analysis of the outcome
PNS,5 there have been few studies concerning the using ordinal logistic regression. The outcome variable was de-
CSF tau levels in PNS disorders.4 In this study, we fined in the three ordered categories (0, the Hughes grade at 6
measured the CSF tau levels in patients with GBS to months was 0; 1, it was I; 2, it was between II and VI). Indepen-
clarify whether it may reflect the degree of axonal dent variables for this model included age at onset, time to nadir,
need for ventilatory support, axonal pattern, CSF tau levels, and
damage and could predict the clinical outcome. timing of when the CSF specimens were acquired. Results were
classed as significant if p ⬍ 0.05.
Methods. Patients. We reviewed medical records of consecu-
tive patients who were admitted to our hospitals between April
Results. The age at onset, sex, time to nadir, Hughes
1998 and March 2004. Twenty-six patients (17 men, mean age
51.4 years) fulfilled the criteria for GBS,6 all of whom were re- grade at nadir, choice of initial treatments, and CSF pro-
cruited to the present study. They were evaluated at the nadir tein levels were not significantly different between the
and 6 months after the onset according to the Hughes functional good and poor outcome groups. On the other hand, the
grading scale,7 and were divided into two groups based on the incidence of diarrhea and axonal pattern on the electrodi-
Hughes grade at 6 months. When the Hughes grade at 6 months
was 0 or I, the cases were classified into the good outcome group agnostic criteria showed a higher tendency in the poor
(n ⫽ 20). The other cases whose Hughes grades at 6 months were outcome group, but the difference was not significant. The
incidence of necessity for ventilatory support and the CSF
tau levels were higher in the poor outcome group (table 1,
figure). Anti-GalNAc-GD1a or GM1 antibodies were posi-
From the Department of Neurology (K.J., S.S., A.O., H.N.), Kohnan Hospi- tive in five of six patients with high CSF tau levels.
tal; Department of Neurology (K.J.), Iwate National Hospital; Departments The univariate model of ordinal logistic regression anal-
of Neurology (A.T., Y.S., Y.I.) and Geriatric and Complementary Medicine
(H.A.), Center for Asian Traditional Medicine Research, Tohoku University
ysis showed that the poor outcome in GBS was associated
Graduate School of Medicine; Department of Neurology (S.K.), Kinki Uni- with necessity for ventilatory support (p ⬍ 0.05), axonal
versity School of Medicine; and Institute of Industrial Ecological Sciences pattern (p ⬍ 0.05), and higher CSF tau levels (p ⬍ 0.01), but
(M.I.), University of Occupational and Environmental Health, Japan. not with older age at onset and longer time to nadir. The
Disclosure: The authors report no conflicts of interest. multivariate model revealed that the poor outcome in GBS
was only associated with higher CSF tau levels (table 2).
Received January 20, 2006. Accepted in final form June 27, 2006.

Address correspondence and reprint requests to Dr. Takeda, Department

Discussion. We showed here that the CSF tau lev-
of Neurology, Tohoku University Graduate School of Medicine, 1-1,
Seiryo-machi, Aoba-ku, Sendai 980-8574, Japan; e-mail: atakeda@ els were significantly higher in GBS cases with poor
em.neurol.med.tohoku.ac.jp outcomes and could be used as an independent prog-
1470 Copyright © 2006 by AAN Enterprises, Inc.
Table 1 Comparison of clinical and laboratory features between
the good and poor outcome groups of patients with Guillain-Barré
syndrome

Good Poor
outcome, outcome, p
n ⫽ 20 n⫽6 Value

Age at onset, y 47.6 ⫾ 19.9 64.2 ⫾ 19.2 0.078

M/F 15:5 2:4 0.084*
Diarrhea, % 20.0 66.7 0.051
Time to nadir, d 7.6 ⫾ 5.0 5.8 ⫾ 1.7 0.46
Need for ventilatory 5.0 50.0 0.028
support, %
Hughes grade at nadir 3.5 ⫾ 0.8 4.2 ⫾ 0.8† 0.11
Initial treatment, %
IV immunoglobulins 85.0 83.3 0.68
Plasmapheresis 10.0 16.7 0.88
Axonal pattern‡ 20.0 66.7 0.051
Protein values in CSF, 85.3 ⫾ 79.1 46.7 ⫾ 17.3 0.24*
mg/mL
Figure. Comparison of tau protein (tau) levels in CSF be-
Tau protein values in 159.6 ⫾ 67.4 341.7 ⫾ 44.5 0.00026
tween the good outcome group (n ⫽ 20) and the poor out-
CSF, ng/mL
come group (n ⫽ 6) in patients with Guillain-Barré
Timing of when CSF 7.2 ⫾ 6.2 7.5 ⫾ 8.0 0.85 syndrome. The functional grading was evaluated at 6
specimens were months after the onset according to the Hughes scale as
acquired, d
follows: 0, healthy; I, minor symptoms or signs, able to
Values are means ⫾ SD where applicable. run; II, ambulates independently; III, able to walk 5
meters with assistance; IV, bed bound; V, requiring as-
* Though not significant, a male predominance and higher CSF sisted ventilation; VI, dead.7 When the Hughes grade at 6
protein levels in the good outcome group were observed, possi- months was 0 or I, the cases were classified into the good
bly due to the small scale of this study. It was well-established
outcome group (n ⫽ 20). Other cases whose Hughes grades
that sex and CSF protein levels do not influence the outcome of
at 6 months were between II and VI were classified into
patients with GBS.1
† One patient was excluded because she died 7 days after the on- the poor outcome group (n ⫽ 6). The mean values of tau
set (n ⫽ 5). are indicated with horizontal lines. The poor outcome
‡ Axonal pattern included the primary axonal and inexcitable group showed higher tau levels than those in the good out-
groups of Hadden’s electrodiagnostic criteria.8 come group (p ⬍ 0.0005). All of the six poor outcome cases
showed high levels of CSF tau (ⱖ280 pg/mL). In contrast,
none of the good outcome cases showed high levels of CSF
tau.
nostic marker for GBS. In contrast with previous
reports,1 neither age nor time to nadir were associ- small study and the clinical features of each case
ated with poor clinical courses in our study, which were not fully described.
could be due to the smaller size of the present study. Tau is a phosphorylated microtubule-associated
Although the univariate model of logistic regression protein primarily localized in neuronal axons and it
analysis showed that the necessity for ventilatory is present not only in the CNS but also in the PNS.5
support and axonal pattern were associated with It has been established that the concentration of
poor outcomes in GBS, the multivariate model anal- CSF tau can reflect the degree of neuronal damage
yses failed to show significance (table 2). Thus, the in CNS disorders such as AD, Creutzfeldt-Jakob dis-
high CSF tau levels were much more closely associ- ease, encephalitis, stroke, and active MS.2-4 The
ated with a poor prognosis of GBS. Because the CSF present results suggest that CSF tau is associated
tau level is known to increase gradually with age,3 with the degree of PNS damage. Because some GBS
the higher levels of CSF tau in the poor outcome cases were reported to be complicated with CNS dys-
group might have been influenced by the older age. functions as shown in Miller Fisher syndrome, the
However, the higher CSF tau levels were indepen- CSF tau might be derived from the putative CNS
dently associated with poor outcomes, even after ad- damage in GBS. Moreover, incidental complications
justment for age (table 2). Thus, CSF tau appears to of subclinical AD might account for the elevated CSF
be a useful and reliable prognostic marker for GBS. tau in poor outcome groups. However, it is plausible
Concerning the CSF tau levels in patients with PNS that the CSF tau in GBS was derived from PNS
disorders, only one article showed that the tau levels lesions, because no cases in the poor outcome group
were low in five patients with GBS,4 but this was a showed either clinical or neuroimaging signs of CNS
October (2 of 2) 2006 NEUROLOGY 67 1471
Table 2 Ordinal logistic regression analysis of factors associated with poor outcome in patients with Guillain-Barré syndrome (n ⫽ 26)

Variable Coefficient Standard error Odds ratio (95% CI) p Value

Age at onset, y 0.080 0.054 1.08 (0.97–1.20) 0.14

Time to nadir, d ⫺0.12 0.20 0.89 (0.59–1.32) 0.55
Need for ventilatory support ⫺5.01 2.60 0.01 (0.000041–1.09) 0.054
Axonal pattern ⫺1.90 2.06 0.15 (0.0026–8.50) 0.36
Tau protein values in CSF, ng/mL ⫺0.039 0.016 0.96 (0.93–0.99) 0.015
Timing of when CSF specimens were acquired, d ⫺0.076 0.18 0.93 (0.65–1.32) 0.67

involvement during the at least 1 year of observa- References

tion. On the other hand, 66.7% and 83.3% of the 1. The Italian Guillain-Barré syndrome Group. The prognosis and main
patients with higher CSF tau levels showed an ax- prognostic indicators of Guillain-Barré syndrome. A multicentre pro-
spective study of 297 patients. Brain 1996;119:2053–2061.
onal pattern (table 1) and positive anti-GalNAc- 2. Vandermeeren M, Mercken M, Vanmechelen E, et al. Detection of ␶
GD1a or GM1 antibodies, respectively. These results proteins in normal and Alzheimer’s disease cerebrospinal fluid with a
sensitive sandwich enzyme-linked immunosorbent assay. J Neurochem
suggest that the CSF tau levels in GBS are closely 1993;61:1828–1834.
associated with axonal injury. Thus, the measure- 3. Arai H, Terajima M, Miura M, et al. Tau in cerebrospinal fluid: a
potential diagnostic marker in Alzheimer’s disease. Ann Neurol 1995;
ment of CSF tau appeared to be useful for estimating 38:649–652.
the presence of axonal degeneration, even in the 4. Süssmuth SD, Reiber H, Tumani H. Tau protein in cerebrospinal fluid
early phase of GBS. Most of the tau detected in CSF (CSF): a blood-CSF barrier related evaluation in patients with various
neurological diseases. Neurosci Lett 2001;300:95–98.
is a proteolytic product of the N-terminal peptide,10 5. Couchie D, Mavilia C, Georgieff IS, et al. Primary structure of high
and therefore we cannot determine which form of tau molecular weight tau present in the peripheral nervous system. Proc
Natl Acad Sci USA 1992;89:4378–4381.
is mainly detected using this method. Further stud- 6. Asbury AK, Cornblath DR. Assessment of current diagnostic criteria for
ies are needed to elucidate the mechanisms responsi- Guillain-Barré syndrome. Ann Neurol 1990;27(suppl):S21–24.
7. Hughes RAC, Newsom-Davis JM, Perkins GD. Controlled trial of pred-
ble for the elevated CSF tau in GBS. nisolone in acute polyneuropathy. Lancet 1978;2:750–753.
8. Hadden RDM, Cornblath DR, Hughes RAC, et al. Electrophysiological
classification of Guillain-Barré syndrome: clinical associations and out-
Acknowledgment come. Ann Neurol 1998;44:780–788.
9. Kusunoki S, Chiba A, Kon K, et al. N-acetylgalactosaminyl GD1a is a
The authors thank Ms. Miwako Tachiya and Ms. Masako Tama-
target molecule for serum antibody in Guillain-Barré syndrome. Ann
mizu for measuring the CSF tau levels; Drs. Masashi Nakamura, Neurol 1994;35:570–576.
Takamasa Takagi, Makoto Hasebe, Sumireko Sekiguchi, and Mi- 10. Johnson GV, Seubert P, Cox TM, et al. The tau protein in human
noru Endou for clinical support; and Mr. Brent Bell for reading cerebrospinal fluid in Alzheimer’s disease consists of proteolytically
the manuscript. derived fragments. J Neurochem 1997;68:430–433.

1472 NEUROLOGY 67 October (2 of 2) 2006

 CSF tau protein: A new prognostic marker for Guillain-Barré syndrome
K. Jin, A. Takeda, Y. Shiga, et al.
Neurology 2006;67;1470-1472
DOI 10.1212/01.wnl.0000240119.29939.c7

This information is current as of October 23, 2006

Updated Information & including high resolution figures, can be found at:
Services http://www.neurology.org/content/67/8/1470.full.html

References This article cites 9 articles, 2 of which you can access for free at:
http://www.neurology.org/content/67/8/1470.full.html##ref-list-1
Citations This article has been cited by 1 HighWire-hosted articles:
http://www.neurology.org/content/67/8/1470.full.html##otherarticles
Subspecialty Collections This article, along with others on similar topics, appears in the
following collection(s):
All Immunology
http://www.neurology.org//cgi/collection/all_immunology
Autoimmune diseases
http://www.neurology.org//cgi/collection/autoimmune_diseases
Peripheral neuropathy
http://www.neurology.org//cgi/collection/peripheral_neuropathy
Prognosis
http://www.neurology.org//cgi/collection/prognosis
Errata An erratum has been published regarding this article. Please see next
page or:
http://www.neurology.org/content/68/7/534.2.full.pdf
Permissions & Licensing Information about reproducing this article in parts (figures,tables) or in
its entirety can be found online at:
http://www.neurology.org/misc/about.xhtml#permissions
Reprints Information about ordering reprints can be found online:
http://www.neurology.org/misc/addir.xhtml#reprintsus

Neurology ® is the official journal of the American Academy of Neurology. Published continuously since
1951, it is now a weekly with 48 issues per year. Copyright . All rights reserved. Print ISSN: 0028-3878.
Online ISSN: 1526-632X.
 Clinical/Scientific Notes
Successful treatment of acquired idiopathic
generalized anhidrosis
F. Palm, MD; C. Löser, MD; W. Gronau, MD;
V. Voigtländer, MD, PhD; and A.J. Grau, MD, PhD

Hypohidrosis and anhidrosis can be caused by various diseases.

Diabetes mellitus, Sjögren syndrome, pure autonomic failure,
Fabry disease, Ross syndrome, thyroid dysfunction, paraneo-
plastic autonomic dysfunction, and congenital absence of sweat
glands are possible diagnoses.1 A less common cause of hypo-
hidrosis/anhidrosis is the acquired idiopathic generalized an-
hidrosis (AIGA). Until now, 64 cases of AIGA have been
reported, 62 being Japanese.2 We here report a European pa-
tient with AIGA.
Case report. A 39-year-old white man presented with a
6-month history of progressive heat intolerance and lack of
sweating except for the axillary zone and parts of his face. The
patient also reported tachycardia and general fatigue. He had
no history of dry mouth, dry eyes, concomitant sharp pain, or
urticaria.
The patient’s history and family history were unremarkable.
The patient did not take any medications. Physical and neuro-
logic examination including autonomic functional tests (heart
rate variability, Valsalva maneuver, respiratory sinus arrhyth-
mia, tilt table test) were normal. Axon reflex testing was not
performed. The thermoregulatory as well as the pilocarpine
sweating tests showed anhidrosis except for the axillary and
periorbital zones (figure, A).
Cranial MRI, chest radiograph, and various laboratory tests
(including anti-GM1 and anti-GQ1B antibodies, IgE level) re-
vealed normal findings. A skin biopsy specimen from the sternum
showed infiltration of sweat glands by CD3-positive lymphocytes
(figure, C). Acquired idiopathic generalized anhidrosis was diag-
nosed, and methylprednisolone was administered (1,000 mg/day
for 3 days IV followed by tapering oral doses for 2 weeks). One
week after therapy initiation, the patient’s sweat production im-
proved. Two months later, the thermoregulatory sweating test
was normal (figure, B). A repeated skin biopsy revealed no more
CD3-positive lymphocyte infiltration (figure, D).
Discussion. AIGA is an uncommon cause of hypohidrosis/
anhidrosis. Most of the reported cases were from Japan. Clinical
features of AIGA are an acute or sudden onset of generalized
anhidrosis with an onset early in life, the absence of other auto-
nomic dysfunction, and a marked response to glucocorticoids. Con-
comitant sharp pain or cholinergic urticaria and elevated IgE
levels have also been described in the majority of patients. These
features were absent in our patient; however, the marked re- Figure. (A) Thermogegulatory sweating tests on admission
sponse to glucocorticoids and the histopathologic findings strongly (note missing blue color for anhidrotic areas). (B) Thermo-
support the diagnosis of AIGA. regulatory sweating test after steroid treatment (note blue
There are three subgroups of AIGA1: The idiopathic pure sudo-
color for improved sweating, especially of trunk and legs).
motoric failure (IPSF), sudomotoric neuropathy, and sweat gland
failure. Most cases of AIGA seem to represent IPSF, as in the case (C) Skin biopsy (CD3 stain) on admission (note lympho-
of our patient. Typical histopathologic findings here include CD3- cyte infiltrations of the sweat glands). (D) Skin biopsy
positive lymphocyte infiltrations of the sweat glands3 and occlu- (CD3 stain) after steroid treatment.
sion of proximal coiled ducts,4 whereas CD3-positive lymphocyte
infiltration seems to be a hallmark of IPSF. Although the etiology
of AIGA is still unclear, immunologic mechanisms contribute to
the disease. The facts that IPSF as a subgroup of AIGA is associ-
ated with CD3-positive lymphocyte infiltration of sweat glands In cases of progressive hypohidrosis/anhidrosis with no other
and that CD3 plays an important role in the induction of cell- pathologic findings, the diagnosis of AIGA should be considered.
mediated disorders support this hypothesis. This may also explain
In most cases, steroid pulse therapy is effective.
the improvement in sweat function with corticosteroids in 78% of
patients with AIGA.2 A deficit in the muscarinic cholinergic recep- From the Departments of Neurology (F.P., W.G., A.J.G.) and Department
tor in eccrine sweat glands or interference in transmission of of Dermatology (C.L., V.V.), Klinikum Ludwigshafen, Ludwigshafen,
acetylcholine to cholinergic receptors is supposed to be involved in Germany.
the pathogenesis of IPSP. This might explain the persisting sweat Disclosure: The authors report no conflicts of interest.
production of the axilla, as seen in our patient. Sweat glands of
the axilla are apocrine glands and are supposed to be under ad- Received May 4, 2006. Accepted in final form October 24, 2006.
renergic control.2 Another explanation would be the early stage of Address correspondence and reprint requests to Dr. Frederick Palm, Depart-
the disease in our patient, with not all sweat glands already being ment of Neurology, Klinikum Ludwigshafen, Bremserstrasse 79, 67063 Lud-
involved. The possibility that the CD3 cells might be directed to wigshafen, Germany; e-mail: palmf@klilu.de
some parts of the eccrine sweat gland itself has to be considered as
well. Copyright © 2007 by AAN Enterprises, Inc.

532 NEUROLOGY 68 February 13, 2007

References 3. Ando Y, Shinji F, Sakashita N, Uchino M, Anso M. Acquired idiopathic
1. Sato K, Kang WH, Saga K, Sato KT. Biology of sweat glands and their generalized anhidrosis: clinical manifestations and histochemical stud-
disorders. II. Disorders of sweat gland function. J Am Acad Dermatol ies. J Neurol Sci 1995;132:80–83.
1989;20:713–726. 4. Ogino J, Saga K, Kagaya M, Kamada A, Kaneko R, Jimbow K. Idiopathic
2. Nakazato Y, Tamura N, Ohkuma K, Yoshimaru K, Shimazu K. Idio- acquired generalized anhidrosis due to occlusion of proximal coiled ducts.
pathic pure sudomotoric failure. Neurology 2004;63:1476–1480. Br J Dermatol 2004;150:589–593.

Occult celiac disease presenting as epilepsy recommendations, he refused a repeat brain biopsy to further
exclude malignancy. He was instead treated with a GFD and
and MRI changes that responded to antiepileptic medication. He has been strictly compliant with
gluten-free diet the diet and has been seizure-free for nearly 2 years, although
the hemiparesis and cognitive difficulties persist. Abnormal
Elizabeth Harper, MD; Harold Moses, MD; and antibody levels have normalized, as is expected with GFD com-
Andre Lagrange, MD, PhD pliance,4 and the contrast-enhancing brain lesions have all
resolved without recurrence (figure).
Celiac disease (CD), an autoimmune disorder involving hyper- Discussion. Our patient with new-onset epilepsy and contrast-
sensitivity to gluten, has been associated with many neurologic
enhancing brain lesions meets the serologic and pathologic diag-
manifestations, most commonly ataxia and neuropathy.1-3 We
nostic criteria for CD. Furthermore, initiation of a GFD resulted
report contrast-enhancing brain lesions and epilepsy in a pa-
in the clinical and radiologic resolution of the brain lesions, sug-
tient with previously occult CD that responded to a gluten-free
gesting that the neurologic symptoms and CD were causally
diet (GFD).
related.
Case report. A 30-year-old, previously healthy man presented
Gastrointestinal involvement of CD may produce malabsorp-
with a 2-year history of headaches and refractory seizures with
tion and some of the neurologic manifestations of CD may be
postictal right hemiparesis and aphasia. He also had a constant
due to vitamin deficiencies.5 However, these data have not been
milder hemiparesis. A brain MRI revealed multiple contrast-
reproducible or conclusive. Moreover, the patient appeared well
enhancing lesions involving white and gray matter. Serologic
testing was unrevealing. A brain biopsy showed inflammation nourished, and serum albumin, hemoglobin, and vitamins E
with reactive gliosis but no microglial nodules. There was also and B12 were all normal. To our knowledge, there is only one
endothelial proliferation, without vasculitis. Special stains for other report of relapsing and remitting contrast-enhancing MRI
periodic acid-Schiff, Gomori methenamine silver, Steiner, her- lesions associated with CD.2 However, that patient had exclu-
pes simplex virus 1 and 2, amoeba, and encephalitis panel were sively white matter lesions, also had oligoclonal bands in the
all negative, as were CSF cultures for viral, bacterial, and CSF, and responded to glucocorticoids but not a GFD. It is
fungal organisms. Despite treatment with glucocorticoids and therefore difficult to distinguish those findings from coincident
multiple antiepileptic drugs, his seizures persisted, and subse- multiple sclerosis in a patient with CD. In contrast, our pa-
quent MRI over the following 10 months found the appearance tient’s cortical involvement and seizures would be distinctly
and disappearance of new lesions. He then presented to our unusual for multiple sclerosis. The protracted, nearly year-long
clinic for a second opinion. Past medical, family, and social course would argue against an alternative diagnosis of acute
history were noncontributory; a systems review revealed no disseminated encephalomyelitis. Finally, the contrast-
fever or weight loss, but he did have chronic constipation and enhancing white and gray matter lesions are not consistent
rash, which were exacerbated during times of increased sei- with MRI changes due to frequent seizures.
zures. Physical examination showed a well-nourished man with Several series have associated CD with epilepsy,3,5 including a
excoriated erythematous papules over his elbows, knees, and child with CD whose refractory epilepsy responded to a GFD.6
buttocks; his general medical examination was otherwise unre- Nonetheless, other than those rare CD cases involving cerebral
markable. Neurologic examination revealed word-finding diffi- calcification, the potential link between CD and epilepsy remains
culty and right-sided weakness. An extensive evaluation for elusive. Moreover, the pathophysiologic basis of neurologic disease
infectious, neoplastic, and inflammatory etiologies was nega- and CD, if one exists, is unknown. However, recent work7 has
tive. Notable negative studies included HIV, vitamin B12 and E found deposition of anti-transglutaminase antibodies in the cere-
levels, screens for collagen vascular disease, serum and CSF bral vessels and brain tissue of patients with gluten ataxia,
angiotensin-converting enzyme levels, CSF cell counts, oligo- thereby suggesting that the antibodies themselves may contribute
clonal banding, IgG index, cytology, and flow cytometry. CT to the neurologic complications of CD.
angiogram showed no evidence for vasculitis or cerebral calcifi- Our case of GFD-responsive CD, epilepsy, and brain MRI le-
cations. Although he denied diarrhea, the patient’s gastrointes- sions supports the idea that CD may involve the CNS and ex-
tinal complaints and rash prompted an evaluation for celiac pands the repertoire of possible neurologic complications
disease. Antigliadin IgG and IgA antibodies as well as anti- associated with CD. Further investigation into the relationship
transglutaminase and anti-endomysial antibodies were all between CD and epilepsy is warranted.
markedly elevated. Small bowel biopsy was diagnostic of CD, From the Department of Neurology, Vanderbilt University, Nashville, TN.
and a skin biopsy was consistent with dermatitis herpeti-
formis. Staining of his previous cerebral biopsy for anti- Disclosure: The authors report no conflicts of interest.
transglutaminase antibodies was not available. Despite our Received May 8, 2006. Accepted in final form October 25, 2006.

Figure. (A) T1 image showing contrast-

enhancing gray and white matter
lesions. (B) T1 image 2 years later,
showing new contrast-enhancing
lesions. (C) T1 image after 12 months
on gluten-free diet, showing resolution
of prior contrast-enhancing regions and
no new lesions.

February 13, 2007 NEUROLOGY 68 533

Address correspondence and reprint requests to Dr. A. Lagrange, Vanderbilt 3. Green PH. The many faces of celiac disease: clinical presentation of
University Medical Center, 6140 Medical Research Building III, 465 21 Ave. celiac disease in the adult population. Gastroenterology 2005;128:S74–
S., Nashville, TN 37232-8552; e-mail: andre.h.lagrange@vanderbilt.edu S78.
4. Di Domenico MR, Annaluisa S, Pluvio R, Iovine C. The role of anti-
endomysium and anti-transglutaminase antibodies in the diagnosis and
Copyright © 2007 by AAN Enterprises, Inc.
follow up of celiac disease. Pediatr Med Chir 2002;24:208–212.
5. Green PH, Alaedini A, Sander HW, Brannagan TH, Latov N, Chin RL.
References Mechanisms underlying celiac disease and its neurological manifesta-
tions. Cell Mol Life Sci 2005;62:791–799.
1. Beyenburg S, Scheid B, Deckert-Schluter M, Lagreze HL. Chronic pro- 6. Mavroudi A, Karatza E, Papastavrou T, Panteliadis C, Spiroglou K.
gressive leukoencephalopathy in adult celiac disease. Neurology 1998;50: Successful treatment of epilepsy and celiac disease with a gluten free
820–822. diet. Pediatr Neurol 2005;33:292–295.
2. Ghezzi A, Filippi M, Falini A, Zaffaroni M. Cerebral involvement in 7. Hadjivassilou M, Grunewald RA, Davies-Jones GAB. Autoantibody tar-
celiac disease: a serial MRI study in a patient with brainstem and cere- geting of brain and intestinal transglutaminase in gluten ataxia. Neurol-
bellar symptoms. Neurology 1997;49:1447–1450. ogy 2006;66:373–377.

Correction

There is nothing staid about STARD: Progress in the reporting of diagnostic accuracy studies

In the editorial “There is nothing staid about STARD: Progress in the reporting of diagnostic accuracy studies” by Karen C.
Johnston and Robert G. Holloway (Neurology 2006;67:740 –741), accompanying the article “The quality of diagnostic accuracy
studies since the STARD statement: Has it improved?” by Smidt et al. (Neurology 2006;67:792–797), the pre-STARD vs post-
STARD comparison was stated to be significant but the data provided demonstrated a statistically insignificant difference. The
authors wish to correct the statement and apologize for the error.

Correction

CSF tau protein: A new prognostic marker for Guillain-Barré syndrome

In the Brief Communication “CSF tau protein: A new prognostic marker for Guillain-Barré syndrome” by K. Jin et al.
(Neurology 2006;67:1470 –1472), the units of CSF tau protein should be pg/mL instead of ng/mL in table 1, table 2, and figure.
The authors regret these errors.

Correction

Correspondence: Neuropsychological deficits in long-term frequent cannabis users

In the reply from authors Lambros Messinis and Panagiotis Papathanasopoulos in the Correspondence concerning “Neuropsycholog-
ical deficits in long-term frequent cannabis users” (Neurology 2006;67:1902), the second respondent’s first name and surname are
transposed. The author’s name is Panagiotis Papathanasopoulos.

534 NEUROLOGY 68 February 13, 2007