Journalof
J Neural Transm [P-D Sect] (1994) 8:131-137

Neural
Transmission
9 Springer-Verlag 1994
Printed in Austria

Marchiafava-Bignami disease with dementia:
severe cerebral metabolic depression revealed by PET

Case Report
S. Pappata 1,2, H. Chabriat 1, M. Levasseur 1, F. Legault-Demare 1,
and J .C. Baron ~,3
Service Hospitalier Fr6d6ric Joliot (SHFJ), and 2 INSERM U334, SHFJ, CEA,
DRIPP, Orsay, and 3INSERM U320, Caen, France
Accepted July 21, 1994

Summary. The Cerebral Metabolic Rate of Glucose (CMRGlu) was measured
with positron emission tomography and 18F-FDG in a patient with MarchiafavaBignami Disease (MBD)-related dementia. Despite MRI evidence of lesions
essentially limited to the corpus callosum (CC), but consistent with the cognitive pattern of cortical dementia, the CMRGlu was markedly reduced in the
frontal and temporo-parieto-occipital association cortices.
Disruption of cortico-cortical networks crossing the CC presumably contributed to, but may not in and by itself explain, the severity of the clinicalmetabolic findings in this patient. An additional role could be played by
microscopic white matter lesions and/or neocortical neuronal loss, which have
been occasionnally observed in post-mortem studies of MBD patients.
Keywords: Marchiafava-Bignami disease, corpus callosum, brain glucose utilization, Positron Emission Tomography, alcoholism.

Introduction
Marchiafava-Bignami Disease (MBD) is a rare complication of chronic severe
alcoholism. Its pathological hallmark is damage of the corpus callosum (CC)
(Castaigne et al., 1971). The advent of CT and MRI imaging has allowed a
better delineation of clinical presentations of MBD other than the classic
letal form, notably a "benign" form with permanent cognitive impairment or
even dementia (superimposed on classic disconnection syndromes) despite
apparently limited CC lesions (Canaple et al., 1992; Nicoli et al., 1994). This
contrast between residual cognitive impairment and apparently limited CC
lesions is intriguing and raises the issue of the associated damage to the cortex
and hemispheric white matter that has been occasionnally reported post-

132

S. Pappata et al.

m o r t e m ( C a s t a i g n e et al., 1971) b u t rarely b y M R I (Nicoli et al., 1994); t h o u g h
n o t m u t u a l l y exclusive, a role for strategically l o c a t e d lesions of the CC,
globally disrupting cognitive function, d o e s h o w e v e r r e m a i n an intriguing
possibility.
This r e p o r t is t h e first to p r e s e n t c e r e b r a l glucose utilization ( C M R G l u )
d a t a in a p a t i e n t with M B D - r e l a t e d d e m e n t i a . Its interest lies in t h e d e m o n stration o f a striking d i s r u p t i o n of synaptic activity in the entire association
n e o c o r t e x , w h i c h was c o h e r e n t with the clinical p r e s e n t a t i o n o f d e m e n t i a b u t
less e x p e c t e d in view of the limited chronic-stage C C lesions r e v e a l e d b y t h e
M R I . A p r e l i m i n a r y a c c o u n t of this w o r k has b e e n p r e s e n t e d earlier ( P a p p a t a
et al., 1992).

Case report
The patient was a 39 year-old unemployed, right-handed male with a long history of
severe alcoholism. He underwent a post-traumatic amputation of the right hand at age 17
years. He was admitted to the hospital three days after sudden onset of confusion and
behavioral changes, without gait disturbances. His family stated that he had been
cognitively normal prior to this episode. At admission, clinical examination showed a
severe global mental deterioration with language and memory disturbances but without
alteration of consciousness, delirium or tremor. The patient showed major impairment
with naming objects to confrontation, word-finding in spontaneous speech, verbal comprehension (both oral and written), and writing. He was disoriented for both time and
space and had a severe anterograde and retrograde amnesia. The rest of the neurological
examination, including gait, was normal apart for absent ankle jerks bilaterally; at careful
examination of the external ocular movements, there was no gaze palsy and no nystagmus
was elicited. The standard laboratory and cerebro-spinal fluid data were normal. The
admission CT Scan showed a clea>cut hypodense lesion of the splenium of the CC,
consistent with the diagnosis of MBD; there was also a generalized, moderate widening of
the cerebral sulci. The patient was treated with detoxification by withdrawal of alcohol,
parenteral B1-B6 vitamins and a combination of meprobamate and haloperidol to prevent anxiety and night-time agitation. He underwent a PET study one month later. By
that time, he had shown some improvement, but was still demented. In addition to mild
neocortical and superior cerebellar atrophy, the MRI T1 - weighted images showed
several small, and one larger, hypointense areas in the splenium of CC; the thickness of
the CC appeared normal, except perhaps at its anterior 2/3 - posterior 1/3 junction which
appeared thinned (Fig. la,b). However, in the mid-sagittal cut, there was also a faintly
visible slit-like hypointensity running antero-posteriorly within the body of the CC (not
shown). The T2 sequences disclosed more conspicuous splenial lesions, and one small
hyperintense area in the genu (Fig. lc,d). There was no evidence of MRI abnormalities
known to be associated with Wernicke encephalopathy (e.g., increased T2 signals near, or
dilatation of, the third ventricule and the aqueduct, or atrophy of the mammilary bodies)
(Gallucci et al., 1990). Neuropsychological evaluation revealed a Wechsler Adult Intelligence Scale verbal IQ of 62 and a Wechsler Memory Scale score of 53. Compared to
admission time, there was an improvement in language performances; he could name 4
out of 5 objects, and better understand verbal orders and write. The spatioconstructive
abilities were altered with "closing in" (placing lines too close to each other). There was
major frontal lobe dysfunction. The verbal fluency was reduced, and the learning of
sequences of hand gestures according to Luria was altered. He showed mental slowness,
abulia and imitation behavior. Severe dementia and amputation of the right hand prevented detailed investigation of inter-hemispheric disconnection but there was definite
left apraxia without tactile naming deficit. A PET study was repeated 3 months later. At
that time the patient showed an essentially unimproved cognitive status.

Marchiafava-Bignami disease with dementia

133

Fig. 1. a-d Magnetic Resonance Imaging (GE 1.5 Tesla), and e Parametric PET
CMRglu images, obtained 1 month after onset of dementia, a,b Sagittal MRI views
showing two lesions in splenium of the corpus callosum (arrows) (Partial Saturation T1
weighted images, TR: 600 msec, TE: 20 msec), e, fl Axial MRI views, showing extensive
hyperintense lesions in the splenium (arrows) and one small lesion in the left part of the
genu of the corpus callosum (arrow) (Spin Echo T2 weighted images, TR: 2,000 msec, TE:
80 msec), e PET images of CMRGlu at four selected brain levels (orbito-meatal plane
+10 mm, +25 ram, +55 mm, and +70 ram, from left to right, respectively). The pixel-bypixel parametric CMRglu images are displayed according to a pseudo-colour scale that
ranges from zero (black) to 7 mg/100 g.min (white). They show a bilateral cortical
hypometabolism, more marked in the prefrontal and temporo-parietal-occipital association cortices than in sensori-motor, auditory and visual primary areas (planes 3 and 4).
The cerebellum (planes 1 and 2) and the deep gray matter structures (plane 3) also appear
relatively spared (see Results and Table 1 for quantitative results)
Methods
PET and lSF-fluoro-2-deoxy-D-glucose (18F-FDG) were used to measure the CMRGlu,
according to a procedure detailed elsewhere (Martinot et al., 1990). PET scans were
performed with the LETI TTV01 time-of-flight four-ring seven-slice camera with axial
and lateral resolutions of about 13 ram. The subject's head was positioned in a headholder so that the lowest slice was 10 mm above the orbito-meatal plane. Correction for
attenuation was carried out using a germanium 68-gallium 68 transmission scan. Plasma
lSFDG curves were obtained from twenty-four arterial blood samples collected from the
radial artery over one hour after the i.v. injection of 5 mCi of the radiotracer. PET images
were acquired from 50 to 60 min post injection. Parametric pixel-by-pixel images of
CMRGlu (rag/100 g/rain) were obtained using the autoradiographic method and the FDG
rate constants and lumped constant derived from normal subjects (Martinot et al., 1990);
this procedure is valid, as the ~SFDG in vivo autoradiographic model is not sensitive

134

S. Pappata et al.

to small variations in these constants, i.e. in the absence of acute non-physiological
conditions.
For PET data analysis, the five brain levels passing through the cerebellum, the
orbito-frontal cortex, the basal ganglia, the corona radiata and the centrum semi-ovale,
respectively, were selected (Martinot et al., 1990). Circular regions of interest (ROIs)
were placed over the cortical structures, thalamus, striatum and cerebellum on the right
side and automatically copied on the left side with respect to, and after appropriate
rotation of, the antero-posterior sagittal axis; due to the limited spatial resolution of the
PET device used, the hippocampal area was not assessed in this study. The data from
cortical ROIs were subsequently averaged according to a limited number of anatomical functional brain areas (Martinot et al., 1990). In addition to the absolute CMRglu values
so obtained, we also evaluated the cortical metabolic "distribution pattern", which is a
classical method for characterizing disease processes relative to one another in a more
sensitive way (Rapoport, 1991). To do this, the absolute cortical regional CMRGlu values
were normalized for the occipital cortex values, i.e. the cortical region that showed least
decrease in CMRGlu in our patient (see below) (SaMe et al., 1991; Schapiro et al., 1993).
The patient's results were compared to those of 31 age matched controls (mean • one SD:
37 _+ 13 years). Statistical significance was assessed using the 95% or 99% individual
prediction limits (IPL) calculated for each brain region (both for absolute CMRglc and
for normalized metabolic values) from the control values according to the following
formula:
IPL = mean + SD.t,
with Student's two-tailed t values of 2.042 and 2.750 for p = 0.05 and p = 0.01,
respectively.
Results

The p a r a m e t r i c C M R G l u images (Fig. l e ) displayed a clear-cut global
h y p o m e t a b o l i s m , m o r e m a r k e d in the anterior and posterior association cortices, with relative sparing of the cerebellum, the d e e p gray nuclei and the
p r i m a r y neocortex. T h e C M R G l u in the basal ganglia, the t h a l a m u s and the
c e r e b e l l u m was in the low range b u t did n o t reach statistical significance
relative to controls (4.82, 4.01, and 3.55 mg/100 g.min, respectively, versus
6.05 + 1.29, 5.61 _+ 1.21 and 4.57 + 1.04 mg/100 g.min, respectively).
Table i shows the absolute neocortical m e t a b o l i c values. The C M R g l u was
significantly d e c r e a s e d for the whole n e o c o r t e x ( - 5 0 % of controls, p < 0.05).
All cortical regions were h y p o m e t a b o l i c , with the lateral prefrontal, medial
frontal (including the anterior cingulate gyrus) and posterior associative
(i.e. parieto-temporo-occipital) cortices reaching statistically significant (p <
0.05) m e t a b o l i c depression bilaterally; the sensori-motor, lateral t e m p o r a l
(which includes the p r i m a r y auditory cortex) and the visual cortices were n o t
significantly hypometabolic. T h e visual cortex was least r e d u c e d (65% of
controls).
The analysis of n o r m a l i z e d metabolic values (Table 1) revealed essentially
similar findings, b u t with e n h a n c e d significance. Significant decreases were
f o u n d for the lateral p r e f r o n t a l (p < 0.01), medial frontal, orbito-frontal and
posterior associative cortices (p < 0.05 for each), b u t n o t for the s e n s o r i - m o t o r
and lateral t e m p o r a l cortices.
Essentially u n c h a n g e d m e t a b o l i c alterations were f o u n d at follow-up P E T
study (data n o t shown). W h e n seen at clinical follow-up at six and 10 m o n t h s

Marchiafava-Bignami disease with dementia

135

Table 1. Average CMRGlu (rag/100 g/min) in whole cortex and cortical areas, and

regional relative metabolic values (cortical region/visual occipital cortex)
Structures

Whole cortex
Lateral prefrontal
Medial frontal
Orbito-frontal
Sensori-motor
Posterior associative
Lateral Temporal
Visual

Region/visual cortex

CMRGlu
Controls

Patient

Controls

Patient

5.59 + 1.25
5.87 + 1.37
6.10 + 1.45
5.52 + 1.39
5.63 __ 1.29
5.33 ___1.22
5.17 + 1.13
6.33 + 1.25

2.93*
2.57*
2.96*
2.77
3.01
2.83*
3.06
4.15

NA
0.92 + 0.09
0.96 + 0.09
0.87 _+ 0.09
0.88 _+ 0.08
0.84 _+ 0.07
0.82 _+ 0.06
1

NA
0.62**
0.71"
0.66*
0.73
0.68*
0.74
1

* p < 0.05 and ** p < 0.01, two tailed probability with respect to 31 age-matched
control subjects (see Methods). N A not applicable
after onset of cognitive impairment, the patient remained d e m e n t e d despite
sustained, though mild, improvement.
Discussion

This patient had the rare presentation of protracted dementia of acute onset
with only partial subsequent improvement, together with CT and M R I lesions
of the CC consistent with the diagnosis of M B D (Castaigne et al., 1971; Nicoli
et al., 1994). The PET studies performed one and three months after onset
showed a conspicuous metabolic depression that was global but preferentially
affected the association neocortex bilaterally, relative to primary neocortex
and subcortical structures. This striking distribution of hypometabolism is
quite consistent with the neuropsychological pattern of "cortical dementia"
observed in this patient, including major frontal lobe dysfunction, m e m o r y
impairment, and disturbances of language and spatio-constructive abilities,
and resembles that reported in advanced Alzheimer's disease (Rapoport et
al., 1991) as well as in one d e m e n t e d patient with neuropathologically-proven
Parkinson's disease but without cortical pathology (Schapiro et al., 1993),
indicating it is specific neither for disease nor for lesion topography.
These major changes in brain glucose metabolism found in our patient
stand in striking contrast with the apparently relatively limited callosal lesions. Could a widespread disruption in transcallosal-based cortical function
result from strategically located callosal lesions and underly, at least in part,
this clinical-metabolic presentation? In favour of this idea, the distribution of
cortical hypometabolism in our patient would be consistent with the regional
pattern of callosal connections, which are dense for associative, and sparse for
primary sensory, cortices (Innocenti et al., 1986). In addition, the frontal and
temporo-parieto-occipital transcallosal fibers cross at the level of the genu and
the splenium, respectively (De Lacoste et al., 1985), which bear the brunt of
the damage in this case. Finally, corpus callosotomy in epileptic subjects often
induces a confusional state (Spencer, 1988; O e p e n et al., 1988), and in the

136

S. Pappata et al.

baboon results in a 30% reduction in CMRGlu in the fronto-temporal regions
(Yamaguchi et al., 1990). However, there are two major inconsistencies with
this hypothesis. First, these effects of corpus callosotomy, both in epileptic
subjects and in baboons, are transient (Spencer, 1988; Oepen et al., 1988;
Yamaguchi et al., 1990), as compared to sustained in our case. Second, though
the MRI suggested somewhat extensive CC lesions, the most conspicuous
damage affected the splenium, which would be expected to chronically impair
almost exclusively the specific transcallosal transfer functions (Geschwind et
al., 1962). It is tempting, therefore, to suggest that other lesions, in addition to
CC damage, contributed to the clinical - metabolic picture in this case. Postmortem studies in MBD occasionnally show hemispheric white matter lesions
and/or cortical neuronal loss (Castaigne et al., 1971); although these findings
come from severe MBD cases going to autopsy, it is possible that such lesions
were present in our case but, as reported recently (Nicoli et al., 1994), escaped
detection by chronic-stage MRI (which showed only a mild and diffuse cortical atrophy which presumably antedated the episode, as suggested by the
admission CT-Scan). However, chronic alcohol abuse per se, would not appear
to have played a direct role, as previous PET studies have shown this does
not result in brain metabolic alterations of either the severity or the distribution pattern seen in this patient (Samson et al., 1986; Volkow et al., 1992;
Adams et al., 1993). Likewise, the metabolic pattern in our patient differed
considerably from that reported in Korsakoff's syndrome, which is characterized by relative sparing of the neocortex with thalamic-limbic system
hypometabolism (Heiss et al., 1992); furthermore, neither the M R I findings
nor the cognitive impairment profile were suggestive of Korsakoff's syndrome
in our case.
In conclusion, in this patient with MBD and with MRI evidence of lesions
limited to the CC (apart from associated mild diffuse cortical atrophy), the
dementia was subserved by a striking brain hypometabolism preferentially
affecting the association cortex. The extent to which the CC damage and the
potentially associated white matter and/or neocortical lesions each contribute
to these major Clinical-metabolic alterations remain uncertain. Further experimental studies combining callosal and extra-callosal lesions might shed
light on this issue.
References

Adams KM, Gilman S, Koeppe RA, Kluin KJ, Brunberg JA, Dede D, Berent S, Kroll PD
(1993) Neuropsychological deficits are correlated with frontal hypometabolism in
positron emission tomography studies of older alcoholic patients. Alcohol Clin Exp
Res 17:205-210
Canaple S, Rosa A, Mizon JP (1992) Marchiafava-Bignami disease: hemispheric disconnection, favourable outcome, neuroradiological aspect. Rev Neurol (Paris) 148: 638640
Castaigne R Buge A, Cambier J, Escourolle R, Rancurel G (1971) La maladie de
Marchiafava-Bignami: 6tude anatomoclinique de dix observations. Rev Neurol
(Paris) 125:179-196
De Lacoste MC, Kirkpatrik JB, Ross ED (1985) Topography of the human corpus
callosum. J Neuropathol Exp Neurol 44:578-591

Marchiafava-Bignami disease with dementia

137

Gallucci M, Bozzao A, Splendani A, Maschiocci C, Passariello R (1990) Wernicke
encephalopathy: MR findings in five patients. Am J Neuroradiol 159:1309-1314
Geschwind N, Kaplan E (1962) A human cerebral deconnection syndrome. A preliminary
report. Neurology 12:675-685
Heiss WD, Pawlick G, Holthoff V, Kessler J, Szelies B (1992) PET correlates of normal
and impaired memory function. Cerebrovasc Brain Metab Rev 4:1-27
Innocenti GM (1986) General organization of callosal connections in the cerebral cortex.
In: Jones ED, Peters A (eds) Cerebral cortex, vol 5. Plenum Press, New York
London, pp 291-353
Martinot JL, Allilaire JF, Mazoyer BM, et al (1990) Obsessive-compulsive disorder: a
clinical, neuropsychological and positron emission tomography study. Acta Psychiatr
Scand 82:233-242
Nicoli F, Vion-Dury J, Chave B, Houallah T, Confort-Gouny S, Cozzone PJ, Gastaut JL
(1994) Maladie de Marchiafava-Bignami: Disconnexion interh6misphdrique, syndrome de Balint, 6volution spontandment favorable. Rev Neurol (Paris) 150:157-161
Oepen G, Schulz-Weiling R, Zimmermann R Birg W, Straesser S, Gilsbach J (1988)
Neuropsychological assessment of the transcallosal approach. Eur Arch Psychiatr
Neurol Sci 237:365-375
Pappata S, Chabriat H, Levasseur M, Legault-Demare F, Baron JC (1992) Severe depression of cerebral glucose utilization (CGU) in dementia due to Marchiafava-Bignami
disease (MBD). Neurology 42 [Suppl 3]:294
Rapoport SI (1991) Positron emission tomography in Alzheimer's disease in relation to
disease pathogenesis: a critical review. Cerebrovasc Brain Metab Rev 4:297-335
Samson Y, Baron JC, Feline A, Bories J, Crouzel C (1986) Local cerebral glucose
utilisation in chronic alcoholics: a positron tomographic study. J Neurol Neurosurg
Psychiatry 49:1165-1170
Sawle GV, Brooks D J, Marsden CD, Frackowiak RSJ (1991) Corticobasal degeneration.
A unique pattern of regional cortical oxygen hypometabolism and striatal fluorodopa
uptake demonstrated by positron emission tomography. Brain 114:541-556
Schapiro MB, Pietrini R Grady CL, Ball MJ, DeCarli C, Kumar A, Kaye JA, Haxby JV
(1993) Reductions in parietal and temporal cerebral metabolic rates for glucose are
not specific for Alzheimer's disease. J Neurol Neurosurg Psychiatry 56:859-864
Spencer SS (1988) Corpus callosum section and other disconnection procedures for
medically intractable seizures. Epilepsia 25 [Suppl 2]: $85-$99
Volkow ND, Hitzemann R, Wang GJ, et al (1992) Decreased brain metabolism in neurologically intact healthy alcoholics. Am J Psychiatry 149:1016-1022
Yamaguchi T, Kunimoto M, Pappata S, et al (1990) Effects of anterior corpus callosum
section on cortical glucose utilization in baboons. Brain 113:937-951
Authors' address: Dr. J. C. Baron, INSERM U320, Blvd. H. Becquerel, BP 5229,
F-14074 Caen Cedex, France.
Received April 25, 1994