Background and Purpose The aim of this study was to evaluate the diaschisis phenomenon in patients presenting with lateral medullary infarct (Wallenberg’s syndrome).
Methods We examined all patients admitted between 1991 and 1993. The localization of lesions was evaluated by MRI. Single-photon emission computed tomographic technique was used to assess cerebral blood flow by two methods (133Xe and hexamethylpropyleneamine oxime) on five slices of brain tissue. Flow values were calculated in 11 regions of interest in each cerebral hemisphere and in the cerebellum and were compared with those obtained in 20 control subjects.
Results Three patients had selective lateral medullary infarct: Relative reduction of flow (133Xe) and of tracer uptake (HMPAO) were observed in one patient in the ipsilateral cerebellum and contralateral hemisphere; in two patients, hemispheric flow values were relatively low, without significant asymmetry. Two patients also presented with cerebellar infarct: Flow drop was severe in the ipsilateral cerebellum, and contralateral reduction in the brain hemisphere was observed in both cases.
Conclusions Lateral medullary infarct can be associated with ipsilateral reduction of flow in the cerebellum, but this phenomenon is inconstant. Severe flow drop suggests infarction in the territory of the posterior inferior cerebellar artery. Contralateral hemispheric flow reduction can also be observed. These phenomena of cerebellar and crossed hemispheric diaschisis are probably related to lesions of tracts from the olivary and reticular nuclei.
Crossed cerebellar hypoperfusion has been described in limited infarcts of the upper brain stem, involving the nucleus rubrum or the lateral pons.1 2 3 4 Lateral and posterolateral medullary infarct is the source of Wallenberg’s syndrome and ipsilateral cerebellar ataxia,5 6 7 8 but a previous evaluation of rCBF in such cases failed to demonstrate cerebellar hypoperfusion.4
The aim of this study was to reevaluate this problem in a group of patients presenting with selective lateral medullary infarct, with the hypothesis that ipsilateral flow reduction could be observed in more severe lesions. We also wished to determine whether such an evaluation could differentiate possible regional diaschisis from the consequences of infarct of the PICA.
Subjects and Methods
We evaluated patients admitted to the Neurological Rehabilitation Unit between 1991 and 1993. Inclusion criteria were clinical Wallenberg’s syndrome and the presence of a lateral medullary lesion on MRI, possibly in the PICA territory.9 MRI was performed with the use of an MR-max machine (General Electric), with a supraconducting magnet operating at 0.5 T, with T1 (repetition time, 400 milliseconds; echo time, 12 milliseconds) and T2 (repetition time, 2000 milliseconds; echo time, 100 milliseconds) sequences.
Exclusion criteria were age older than 70 years, previous neurological or psychiatric disorders, alcoholism, vigilance disorders, current neuroleptic or antiepileptic treatment, and infarct in other territories of the brain. In each patient the periventricular hyperintensity on MRI was evaluated with the use of the Fazekas index.10
For CBF and activity analyses we used SPECT, with a Tomomatic 564 apparatus (Medimatic), under two successive conditions: (1) during continuous inhalation of 133Xe and (2) after injection of 555 mBq IV of 99mTc-HMPAO. The HMPAO was administered immediately after the end of the 133Xe study by an intravenous catheter that had been placed before the first investigation to avoid stress and anxiety. The patient’s head was kept in the same position for both evaluations.
The examination was conducted under standard conditions11 12 after a 15-minute adaptation period in dim light and silence. The partial alveolar pressure for CO2 was recorded with a Beckman LB2 capnograph for the 133Xe study.
The Tomomatic 564 defined five 15-mm-thick slices over the OM plane (ie, OM+10+30+50+70+90 mm). The full width at half maximum was 16 mm and the pixel 25 mm2. As in previous studies,10 rCBF and activity values were quantified in the same way (1) in nine areas (size, 650 to 800 mm2; ≥25 pixels) traced on each cerebral hemisphere on the intermediate slice (OM+50 mm) passing through the basal ganglia (including frontal anterointernal, frontal anteroexternal, frontal posteroexternal, temporal anterior, temporal posterior, temporo-occipital, occipital, lenticular, and thalamic) and (2) in the cerebellum on the lower slice (OM+10 mm). Each area was automatically defined by the computerized system and later adjusted manually (blind to clinical data). Mean flow or activity values were later calculated in 4 hemispheric areas: frontal (3 areas), temporal (2 areas), temporo-occipital cortex (2 areas), and subcortical nuclei (2 areas). The reproducibility of the method has been previously described.13
For both methods, we later calculated AI values for the cerebellum [AI%=(healthy side rCBF−lesion side rCBF)×200/(healthy side rCBF+lesion side rCBF)] and hemisphere areas [AI%=(lesion side rCBF−healthy side rCBF)×200/(lesion side rCBF+healthy side rCBF)].
For the 133Xe study, rCBF and AI values of patients were compared with those of 20 control subjects (12 men, 8 women; age range, 22 to 73 years; mean age, 38.1 years). In these control subjects, AI values were not significantly correlated with age. For the HMPAO evaluation, we compared the AI values with those calculated in a group of 6 normal subjects.13
We also performed a qualitative, visual analysis of the cartographic images (without knowing the side of the medullary or cerebellar lesion on MRI) to search for asymmetry in the cerebellum, frontotemporal cortex, temporo-occipital cortex, basal ganglia, and thalamus. Each asymmetry was quantified as clear or severe (++), moderate (+), or mild or absent (0); the interrater reliability coefficient was 0.816.
Five patients were included in the study (Table 1⇓). On MRI, 3 had pure lateral medullary infarct without any lesion in the upper brain stem, the cerebellum, or the cerebral hemispheres. Clinical deficits and parenchymal injury were more severe in the second patient, in whom MRI showed extension of the lesion in the internal medullary structures (Fig 1⇓). Two other patients had cerebellar infarct in the ipsilateral PICA territory. Patient 4 presented with relatively severe periventricular hyperintensity (Fazekas’ grade 2), but this was relatively symmetrical; furthermore, two small lacunar lesions were observed: one in the right anterior capsule and the other in the posterior part of the left posterior capsule. In patient 1 Doppler ultrasonography revealed vertebral artery occlusion on the side of the medullary lesion, which was associated with cervical stenosis of the left carotid artery; this evaluation did not reveal cervical arterial stenosis or thrombosis in the other cases. Vertebral artery thrombosis was confirmed by angiography in patient 1, but flow was preserved in the PICA territory. In patients 2 and 3 angiography did not reveal stenosis of the vertebral artery or of the PICA. Carotid angiography was not performed.
In patients with pure Wallenberg’s syndrome (patients 1, 2, and 3), visual inspection (Table 2⇓) of brain images obtained with 133Xe revealed cerebellar asymmetry in patient 2 (Fig 1⇑), with flow values inferior on the side of the medullary lesion; furthermore, an inverse asymmetry was observed in the cerebral hemisphere. The cerebellar asymmetry was less severe on images obtained with HMPAO, but hemispheric asymmetry was similar. In patients 1 and 3 no cerebellar or hemispheric asymmetry could be observed, but flow values were relatively low on both sides.
These results were confirmed by quantitative analysis of the rCBF study with the use of 133Xe (Table 3⇓). In patient 2 significant AI values were observed in the cerebellum (AI=17.14%) and subcortical nuclei (AI=15.46%), with a tendency toward asymmetry noted in the temporal cortex (AI=7.92%); in patients 1 and 3 rCBF was relatively low compared with values obtained in control subjects but was without significant asymmetry. In the HMPAO study (Table 4⇓), cerebellar AI in patient 2 also reached significance but was lower than that obtained with the use of 133Xe; crossed hemispheric diaschisis was observed in the frontal cortex and subcortical nuclei.
In both patients presenting with cerebellar infarct (patients 4 and 5), visual inspection revealed severe cerebellar asymmetry and crossed hemispheric diaschisis (Fig 2⇓), which was less important in patient 5, who was evaluated 465 days after stroke. On quantitative analysis, cerebellar rCBF drop was severe on the side of the infarct, and inverse asymmetry was observed in the hemispheres, with AI significantly increased in patient 4 in the frontal (AI=22.12%) and temporal (AI=11.19%) cortices and in patient 5 in the subcortical nuclei (AI=19.40%). With HMPAO the temporal AI was increased in patient 4 and the subcortical AI in patient 5.
This study revealed (1) that pure lateral medullary infarct may be associated with relative hypoperfusion in the ipsilateral cerebellum and contralateral hemisphere and discrete hemispheric flow reduction and (2) that cerebellar rCBF reduction is less severe than in the case of PICA infarct.
In contrast to previous findings,4 pure lateral medullary infarct associated with Wallenberg’s syndrome can be the source of ipsilateral cerebellar flow reduction. The inconstancy of this phenomenon requires discussion. In patient 2 it could not be related to arterial stenosis. As suggested by Fazekas et al,4 no correlation could be demonstrated between the presence of clinical cerebellar syndrome and cerebellar hypoperfusion; ataxia was not more severe in this case. The main clinical difference with the two other patients was severe deficit of cranial nerves V, IX, and X with dysphagia and dysphonia, which suggested more caudal and medial diffusion of the lesion8 in the region of the nucleus ambiguus.14 MRI hyperintensity was also more extended in the internal medulla; however, brain stem lesions are relatively difficult to assess with 0.5-T MRI15 and can be more precisely defined with 1.5-T apparatus.8 In more medial medullary areas, three main systems project onto the cerebellar cortex, the injury of which could be responsible for the diaschisis we observed.14 The olivocerebellar tract is probably the most important and constitutes the largest component of the inferior cerebellar peduncle; it produces a powerful excitatory action on Purkinje cells. Reticular nuclei, particularly the paramedian reticular nucleus and the lateral reticular nucleus, send important projections to the cerebellum. Furthermore, the rostral spinocerebellar tract, which maintains a retro-olivary position, enters the cerebellum in the inferior cerebellar peduncle for nearly one third of its fibers.
Cerebellar hypoperfusion was associated with crossed hemispheric diaschisis, similar to that observed in unilateral cerebellar hematomas16 17 and infarcts.18 As previously described, this phenomenon was relatively diffuse but was more severe in the frontotemporal cortex, basal ganglia, and thalamus and spared occipital structures. Such a contralateral effect could be explained by indirect diaschisis via the cerebellothalamocortical pathway, by direct and crossed deactivation, or by both phenomena. Like corticocerebellar diaschisis,19 medulla-hemisphere diaschisis is probably related more to lesions of paths joining the medulla to the thalamus and/or the cortex, such as reticulothalamic, spinothalamic,20 or vestibular-cortical fasciculi, than to injury of fibers linking the cortex to the medulla, eg, corticoreticular fibers arising from the sensorimotor areas.14 21
In patients presenting with cerebellar infarct in the PICA territory, ie, in the inferior and medial part of the cerebellum, the cerebellar rCBF drop was much more severe than the previously described medulla-cerebellum diaschisis and was compatible with neuronal destruction. In patient 4 the contralateral hemispheric hypoperfusion (“crossed hemispheric diaschisis”)16 17 18 was severe and of the same intensity as in some patients with unilateral cerebellar hemorrhage.16 The possibility that this asymmetry could reflect subcortical injury can be discussed. However, lesions did not prevail on the side of the relative flow drop; furthermore, the thalamus, in which a lesion could be responsible for more diffuse cortical diaschisis,22 was spared. In patient 5, in whom rCBF evaluation was performed 465 days after stroke, the diaschisis was discrete in the frontotemporal cortex and more severe in the basal ganglia and thalamus. In both patients flow values were symmetrical in the occipital cortex. These remote effects, the inverse of the more classic corticocerebellar diaschisis,19 can be associated with lesions of both cerebellocortical and medullocortical paths. Our results also suggest that through assessment of rCBF one may differentiate pure Wallenberg’s syndrome from lateral medullary syndrome associated with cerebellar infarct, which is often difficult on clinical grounds.8 23 24 However, SPECT studies were not performed at the acute stage, which could lead to more definite conclusions; further studies are necessary to confirm this particular point.
In two patients (patients 1 and 3), relatively low rCBF values were observed in both hemispheres. In these patients MRI did not reveal periventricular hyperintensity that could be associated with leukoarariosis and reduced flow values. The hypothesis of diffuse arteriopathy cannot be ruled out, even in the younger patient (patient 3). Medullary structures such as the reticular formation have relatively diffuse cortical efferences or afferences, lesions of which could also contribute to this phenomenon.
Our study revealed discrepancies between 133Xe and HMPAO data. Since the data were obtained in the same SPECT session, these discrepancies cannot be due to spontaneous evolution. AI values were in most cases higher with 133Xe, as previously observed in a similar study of normal subjects.13 Such a phenomenon could be related to (1) the septal penetration of photons from technetium, which is able to contaminate activity in regions of interest,25 and (2) the mechanism of HMPAO uptake, which is related to the conversion of lipophilic to hydrophilic complexes, in relation to the metabolic activity of neurons. We also observed the interindividual variability of AI, particularly in patient 4 and in the frontal cortex. Since regions of interest were placed in a very similar manner, these discrepancies cannot be attributed to technical problems. Furthermore, in this particular case visual examination of HMPAO images did not reveal the frontal asymmetry obtained with the use of 133Xe (Fig 2⇑).
Selected Abbreviations and Acronyms
|CBF||=||cerebral blood flow|
|PICA||=||posterior inferior cerebellar artery|
|rCBF||=||regional cerebral blood flow|
|SPECT||=||single-photon emission computed tomography|
- Received January 24, 1995.
- Revision received April 25, 1995.
- Accepted April 27, 1995.
- Copyright © 1995 by American Heart Association
Sakai F, Aoki S, Kan S, Igarashi H, Kanda T, Tazaki Y. Ataxic hemiparesis with reductions of ipsilateral cerebellar blood flow. Stroke. 1986;17:1016-1018.
Fazekas F, Payer F, Valetitsch H, Schmidt R, Flooh E. Brain stem infarction and diaschisis: a SPECT cerebral perfusion study. Stroke. 1993;24:1162-1166.
Wallenberg A. Acute bulbaraffection (Embolie der Art. Cerebellar. Post. Inf. Sinistr.). Arch Psychiatr Nervenkr. 1895;27:505-540.
Currier RD, Giles CL, De Jong RN. Some comments on Wallenberg’s lateral medullary syndrome. Neurology. 1961;1:778-792.
Currier RD. Syndromes of the medulla oblongata. In: Vinken PJ, Bruyn GW, eds. Localization in Clinical Neurology, Handbook of Clinical Neurology, Vol 2. Amsterdam, Netherlands: North Holland Publishing Co; 1969:217-237.
Kim JS, Lee JH, Suh DC, Lee MC. Spectrum of lateral medullary syndrome: correlation between clinical findings and magnetic resonance imaging in 33 subjects. Stroke. 1994;25:1405-1410.
Ross MA, Biller J, Adams HP, Dunn V. Magnetic resonance imaging in Wallenberg’s lateral medullary syndrome. Stroke. 1986;17:542-545.
Fazekas F, Chawluk JB, Alavi A, Hurtig MI, Zimmerman RA. MR abnormality at 1.5 T in Alzheimer dementia and normal aging. Am J Neuroradiol. 1987;8:421-426.
Steinling M, Guerouaou D, Blanc D, Bouchez P, Arnott G, Vergnes R. Mesure du débit sanguin cérébral (local) en tomographie d’émission gamma: valeurs normales et critères quantifiés d’interprétation. Ann Radiol. 1986;29:7-12.
Carpenter MB, Sutin J. Human Neuroanatomy. Baltimore, Md: Williams & Wilkins Co; 1983.
Rousseaux M, Steinling M. Crossed hemispheric diaschisis in unilateral cerebellar lesions. Stroke. 1992;23:511-514.
Baron JC, Bousser MG, Comar D, Castaigne P. Crossed cerebellar diaschisis in human supratentorial brain infarction. Trans Am Neurol Assoc. 1980;8:120-135.
Willis WD. The origin and destination of pathways involved in pain transmission. In: Wall PD, Melzack R, eds. Textbook of Pain. Edinburgh, Scotland: Churchill Livingstone, Inc; 1984:88-99.
Kuypers HGJM. Cortico-bulbar connections to the pons and lower brainstem in man: an anatomical study. Brain. 1958;81:364-388.
Steinling M, Coequyt S, Dubois P, Vergnes R. 123Iodoamphétamine (IAMP): tomographe conventionel ou tomographe spécialisé? Circ Métab Cerveau. 1985;2:277-284.