- Email: [email protected]
Patterns of cerebellar atrophy in patients with chronic epilepsy: a quantitative neuropathological study
Epilepsy Research 41 (2000) 63 – 73 www.elsevier.com/locate/epilepsyres
Patterns of cerebellar atrophy in patients with chronic epilepsy: a quantitative neuropathological study R. Crooks a, T. Mitchell b, M. Thom a,* a
Department of Neuropathology, Institute of Neurology, Uni6ersity College London, Queen Square, London WC1N 3BG, UK b Department of Neurology, Institute of Neurology, Uni6ersity College London, Queen Square, London WC1N 3BG, UK Received 17 August 1999; received in revised form 7 April 2000; accepted 11 April 2000
Abstract Cerebellar atrophy occuring in patients with chronic epilepsy is considered either to be a sequel of cumulative seizure-mediated cell loss or a side effect of phenytoin treatment but there is little neuropathological data regarding the distribution of this cerebellar damage. We aimed to address if there is any relationship between the localisation of the cortical pathology in symptomatic epilepsy and the pattern of neocerebellar atrophy. A quantitative neuropathological post mortem analysis of the lobular distribution of hemispheric cerebellar atrophy in 16 patients with chronic epilepsy and four controls was carried out. Cerebellar atrophy, as measured by significant reductions in hemispheric linear Purkinje cell densities was confirmed in the epilepsy patients (P= 0.015) and even where the cerebellum appeared macroscopically normal, Purkinje cell loss was evident (P = 0.062). Two distinct patterns of atrophy were observed, predominantly involving either the anterior or posterior cerebellar lobes. Posterior lobe atrophy was more often associated with old fronto-temporal contusions and may be post traumatic in aetiology rather than a result of excitotoxic damage mediated via cerebro – cerebellar pathways. As the majority of patients showing either pattern of atrophy had received phenytoin treatment, we concluded that it is unlikely that this drug acts alone in inducing the Purkinje cell loss. © 2000 Elsevier Science B.V. All rights reserved. Keywords: Cerebellar atrophy; Purkinje cell density; Epilepsy
1. Introduction Cerebellar atrophy has long been recognised from both neuropathological and neuroimaging studies to occur in a proportion of patients with chronic epilepsy (Spielmeyer 1930; Dam 1987; Botez et al., 1988). This atrophy has been variably * Corresponding author. Fax: +44-171-9169546. E-mail address: [email protected] (M. Thom)
attributed to seizure mediated cellular damage, a toxic side effect of phenytoin, transneuronal cerebellar degeneration following crossed cerebral hemiatrophy, anoxic-ischaemic injury during seizures or a combination of these factors acting synergistically (Ney et al., 1994; Savic´ and Thorell, 1996; Specht et al., 1997). Despite its long recognition there have been very few pathological studies regarding the anatomical distribution of cerebellar atrophy in epilepsy. The most detailed
0920-1211/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 0 - 1 2 1 1 ( 0 0 ) 0 0 1 3 3 - 9
64
R. Crooks et al. / Epilepsy Research 41 (2000) 63–73
study to date was based on a description of five cases illustrating the atrophy observed with each aforementioned processes (Gessaga and Urich, 1985); for example a ‘post-ictal pattern’ showed atrophy at the level of the horizontal fissure in the vascular watershed territory of the lateral lobes whereas phenytoin toxicity resulted in more diffuse cerebellar atrophy. There is evidence to suggest that cerebellar atrophy may be of clinical importance in patients with epilepsy. It has been associated with a worse prognosis, in terms of seizure control, following temporal lobectomy (Specht et al., 1997). Furthermore, pathology limited to the cerebellum has been associated with a cognitive affective syndrome (Schmahmann and Sherman, 1998) which may also be of relevance in patients with epilepsy (Specht et al., 1997; Pulliainen et al., 1998). In view of the above, we carried out a quantitative neuropathological analysis of the patterns of cerebellar hemispheric atrophy in post mortem material from patients with chronic epilepsy. We aimed to address if there is any relationship between the pattern of cerebellar atrophy and the identified neocortical pathology implicated in the seizure genesis to suggest seizure-mediated cell loss via cerebro–cerebellar pathways.
2. Methods Post mortem cerebellar tissue from 16 patients with epilepsy (age range 21 – 84, mean 49 years) was included in this study. All the post mortems or routine neuropathological examinations had been carried out between the years 1992 and 1998 in the Department of Neuropathology at the National Hospital for Neurology and Neurosurgery, London. The cases were selected to include a range of localised cerebral pathologies relevant to the patients epilepsy disorder (Table 1); these included hippocampal sclerosis (cases 2, 4, 5, 12, 14 and 15) which was bilateral and asymmetrical in two (cases 2 and 4) and associated with cortical hemi-atrophy in two (cases 5 and 15), end folium hippocampal sclerosis (cases 7, 11 and 13), malformations of cortical development including microdysgenesis (cases 1 and 5), grey matter
heterotopias (cases 7 and 14), vascular malformations (cases 3 and 9), glio-neuronal hamartoma (case 4) and early cerebral hypoxic-ischaemic injury with cortical ulegyria (case 11). In addition, four of the older patients had established focal cerebral and lacunar infarcts (cases 4, 7, 10 and 14) and old fronto-temporal contusions from previous head injuries, were identified in six cases (cases 7, 8, 9, 10, 14 and 16). In eight of the 16 cases no cerebellar atrophy was evident on routine macroscopic examination (cases 2, 3, 6, 7, 11, 12, 13 and 14) and this, in most cases where available, correlated with pre-mortem neuroimaging findings (Table 1). The hospital notes from these patients were also reviewed and the data are presented in Table 2. Eleven patients had complex partial seizures, nine with secondary generalised seizures and five patients had generalised tonic clonic seizures. In five patients the seizure frequency was severe, with maximum seizure frequencies of four per week (cases 1, 2, 3, 11 and 14). The EEG data obtained from review of the clinical records showed distinct cortical lateralising foci in seven patients (cases 2, 3, 5, 6, 8, 10 and 11). Despite our attempts to include patients who had not been treated with phenytoin, we were able to obtain neuropathological tissue from only one such patient (case 6). Two patients (cases 9 and 14) had recorded episodes of phenytoin toxicity with an acute reversible cerebellar syndrome. None of the patients had a documented history of alcohol or drug abuse. Similar tissue sections were sampled from the remaining cerebellar tissue of each case; a traditional plane of cut was made at right angles to the folia through the dentate nucleus to include the entire dorsoventral extent of both left and right hemispheres. An ‘anterior’ section corresponded to the anterior lobe, and ‘superior’, ‘horizontal’, ‘inferior’ and ‘tonsil’ sections represented the posterior lobe (Fig. 1). The ‘horizontal’ section included both superior and inferior semilunar lobules and the horizontal fissure and thus represented the vascular watershed area. Twenty micron thick sections were cut and stained with 0.01% cresyl violet from which the total length of the Purkinje cell layer was measured using a Leica
Table 1 Neuropathological findings in the 16 epilepsy patientsa Case
Cerebellum: macroscopic description, fixed hindbrain weight (% total brain weight), where recorded
Histological neuropathological diagnosis relevant to epilepsy
1 2 3
SUDEP SUDEP SUDEP
Atrophy of anterior lobes No atrophy No atrophy
Parietal lobe microdysgenesis Bilateral HCS L\R Right occipital lobe cortical scarring
4
Lobar pneumonia Bronchopneumonia SUDEP Peritonitis
Microcephaly (995) Left hippocampal atrophy (1390) Right occipital cavernoma (operated) (1380) Right occipital infarct (880)
Symmetrical atrophy
Left hippocampal and cortical hemiatrophy (1000) Normal (1160) Right parietal laminar heterotopias, recent left frontal infarct; old fronto temporal contusions (1465) Old frontal contusions (1220) Right frontal contusions (1100) Old fronto-temporal contusions, lacunar infarcts (1310) Left frontal and occipital ulegyria (1030) Right temporal lobe atrophy and arachnoid cyst (950) Left hippocampal atrophy (1650)
Symmetrical atrophy 90 g (9%) No atrophy No atrophy
Left occipital neuronal — glial hamartoma Bilateral HCS R\L Left temporal cortical microdysgensis; left HCS and left cortical hemiatrophy Acute neuronal changes in hippocampus Mild EFS
Atrophic 95 g (7.7%) Atrophy 108 g (9.8%) Atrophy in watershed area 120 g (9.1%)
No specific diagnosis Left amygdala telangiectasia No specific diagnosis
No atrophy No atrophy 115 g (12%)
Mild EFS Right HCS and temporal lobe ulegyria
No atrophy
5 6 7
8 9 10
SUDEP SUDEP Ca lung
11 12
SUDEP SUDEP
13
SUDEP
14
Bronchopneumonia
15
Bronchopneumonia Bronchopneumonia
16
a
Bilateral subependymal heterotopias; old fronto temporal contusions, old left PCA infarct (990) Left cerebral hemiatrophy (1180)
No atrophy 145 g (14%)
Mild left EFS; acute neuronal changes in subiculum Right HCS
Mild atrophy 120 g (10%)
Left HCS and left cortical hemiatrophy
Old contusions (975)
Atrophy, 100 g (10.2%)
No HCS
R. Crooks et al. / Epilepsy Research 41 (2000) 63–73
Cause of death Macroscopic appearance of brain (fixed brain weight (g))
SUDEP, sudden and unexpected death in epilepsy; PCA, posterior cerebral artery; HCS, hippocampal sclerosis; EFS, end folium sclerosis.
65
Case
Seizure history: duration, frequency (max.), type
1
21 (F)
20 years, 6-7/day; FC, CPS, SGS, MJ
2
21 (F)
3
23 (F)
4
76 (F)
5
59 (F)
6 7
27 (F) 70 (F)
8
44 (M)
9
47 (M)
10
69 (M)
11
40 (F)
12
67 (F)
13
27 (F)
14
49 (F)
15 16
73 (F) 84 (M)
EEG
Abnormal irregular slow waves and bilateral sharp waves and spikes. 3 years, 3–4/week; CPS, Generalized slow and SGS, SE sharp waves over left hemipshere 15 years, 2–3/day; MJ, Frequent paroxysmal CPS, SGS epileptiform discharges over right hemipshere 76 years, 7/year; CPS, SGS Irregular slow waves over both temporal lobes 46 years, 1/month; SE, Left sided focus CPS, SGS 12 years, 1/week; GTCS Left fronto-temporal focus 51 years, 1/year; CPS, SGS Irregular slow waves left temporal; right temporal spikes 37 years, 2–3/year; CPS, Right sided focus SGS 45 years, ‘frequent’; GTCS Abnormal irregular slow waves; bilateral sharp waves and spikes 62 years; CPS Right temporo-parietal focus 32 years, 18/month; CPS, Left temporal focus SGS 64 years, 8–12/month; Bilateral central slow wave GTCS disturbance 16 years, 1/week; MJ, Abnormal activity both GTCS, PS L&R fronto-temporal lobes 20 years, 6–7/day; FC, Abnormal irregular slow CPS, SGS, MJ waves and bilateral sharp waves and spikes. 72 years, 1–2 month; CPS Within normal limits 68 years, 1/month; GTCS Generalized abnormalities
Neuroimaging
AED treatments
Duration (maximal dose) of Phenytoin treatment when documented
CT normal
PT, CBZ, PhB, SVP, Cl, Lam
B5 years (5–100 mg)
MRI – left HCS
PT, CBZ, Vig, Cl, SVP, 10 months (600 mg) Lam
MRI – atrophy right occipital lobe
PT, PhB, Cl, Lam, Vig, NA CBZ, SVP
CT – cerebral and cerebellar atrophy CT – left hemiatrophy and cerebellar atrophy Not done MRI – parietal lobe infarct & old frontal contusions. CT – cerebral and cerebellar atrophy. MRI – minor cerebellar atrophy
PT, PhB
B5 years
PT, SVP, CBZ, PhB, Cl CBZ, SVP PT, CBZ, PhB, Cl
B10 years No PHT NA
PT, SVP, PhB, CBZ, 23 years (150 mg) Cl PT, CBZ, PhB, ViG, Cl 2 years
CT – normal
PT, CBZ, PhB, Lam
B5 years
CT – left hemiatrophy
19 years (300 mg)
Not done
PT, CBZ, PhB, Vig, Lam, Acet PT, PhB
NA
CT-atrophy
PT, PhB, CBZ, SVP
11 months (300 mg)
CT-Normal
PT, CBZ, Lam, Cl, Vig 9 years
Not done CT – cerebral and cerebellar atrophy
PT, CBZ, Cl, PhB PT, PhB
NA 20 years+
a FC, febrile convulsion; CPS, complex partial seizures; MJ, myoclonic jerks; GTCS, generalised tonic clonic seizures; SGS, secondary generalised seizures; HCS, hippocampal sclerosis; PT, phenytoin; CBZ, carbamazepine; PhB, phenobarbitone; SVP, sodium valproate; Cl, Clonazepam; Lam, Lamotrigine; Vig, Vigabatrine; Acet, acetazolamide; NA, information not available.
R. Crooks et al. / Epilepsy Research 41 (2000) 63–73
Age (sex)
66
Table 2 Clinical data of the 16 patients with chronic epilepsya
R. Crooks et al. / Epilepsy Research 41 (2000) 63–73
Q500 image analyzer (Leica Imaging Systems, Cambridge) by a single observer (RC). The total number of Purkinje cells (defined for our study as large nucleolated neurones lying at the junctions of the molecular and granule cell layers with prominent apical dendrites) were counted at a magnification of × 20 on a Leica DM RB microscope by two independent observers (RC and MT) with good reproducibility. Linear Purkinje cell densities/mm (PCD) were then calculated for each region of each hemisphere. Mean PCD values between anatomical regions and patient groups were compared using an independent Ttest and Pearson’s correlation coefficient was used to assess for any linear associations. Identical examination and quantitative analysis was carried out on four neurologically normal controls of ages 27, 47, 52 and 56 years (mean age 45 years). The cause of death in the control cases included a road traffic accident, aplastic anaemia, bronchop-
67
Fig. 2. Histogram of the mean left and right cerebellar hemispheric Purkinje cell densities in each subregion in controls, all epilepsy patients and in the two epilepsy groups; group 1 (cases 1 – 5) showing predominant anterior lobe atrophy and group 2 (cases 6 – 16) posterior lobe atrophy.
neumonia with pulmonary fibrosis and bronchial carcinoma respectively. In none of these cases was significant neuropathological or neurodegenerative disease identified.
3. Results
Fig. 1. Figure illustrating the five areas from each cerebellar hemisphere at the level of the dentate nucleus referred to in the text & tables. (1) ‘Anterior section’ comprises the central and anterior quadrangular lobule (Anterior lobe of the cerebellum). (2) ‘Superior section’ comprises the posterior quadrangular lobule and the anterior half of the superior semilunar lobule and includes the posterior superior fissure. (3) ‘Horizontal section’ comprises the posterior part of the superior semilunar lobule and the entire inferior semilunar lobule and includes the horizontal fissure. (4) ‘Inferior section’ comprises the gracile and biventer lobules and the prepyramidal fissure. (5) ‘Tonsil section’ and comprises the cerebellar tonsil. (2–5, Posterior lobe of cerebellum).
Histological examination revealed various degrees of cerebellar atrophy with Purkinje cell loss, preservation of basket cells, Bergmann gliosis and focal granule cell depletion in all the epilepsy patients. In some cases torpedo like axon swellings were present on residual Purkinje cells as evidence of ongoing cell loss. Quantitative analysis confirmed this finding with significant reductions in Purkinje cell linear densities (PCD) in the epilepsy group 3.2/mm (SD 2.1) compared to controls 6.14/mm (SD 0.19, P= 0.015). In eight patients without macroscopically apparent cerebellar atrophy, lower mean PCD were also observed (4.5/mm, P= 0.062). In the control group, slightly higher PCD were recorded in the anterior, superior and tonsil sections than in the horizontal
68
R. Crooks et al. / Epilepsy Research 41 (2000) 63–73
and inferior sections (Table 3; Fig. 2) which reflects the normal anatomical variation in Purkinje cell densities (Hall et al., 1975). Two distinct patterns of cerebellar atrophy were observed in the epilepsy patients. The first, (group 1, cases 1–5) showed a more significant reduction in PCD compared to controls in the anterior and superior sections (Table 3, Fig. 2). A second pattern (group 2, cases 6 – 16) showed greater reduction in PCD in the horizontal, inferior and tonsil sections. In cases with more severe atrophy, these two distinct patterns were obvious from macroscopic examination (Fig. 3). In two patients (cases 5 and 15) cerebral hemiatrophy with hippocampal sclerosis was present and although crossed cerebellar atrophy was not obvious on macroscopic or histological examination alone, quantitation revealed a greater reduction in PCD in the contralateral than ipsilateral cerebellar hemisphere in both cases. In six patients (cases 2, 3, 6, 8, 10 and 11), a well defined lateralising EEG focus was identified (Table 2) and in three of these (cases 2, 3 and 11) this correlated with an underlying unilateral neocortical pathology. In two patients (cases 4 and 12) a unilateral cortical pathology or malformation was identified and considered likely to be the cause of the patients seizures. In these ten patients with lateralising cortical lesions, seven (cases 2, 4, 5, 8, 10, 12 and 15) showed reduced mean PCD in the contralateral than ipsilateral cerebellar hemisphere, but the difference was not significant (P = 0.13). A trend of lower mean PCD with increasing age of the patient was observed but there was no significant correlation (P =0.107). Also we observed a trend of lower mean PCD in patients with longer seizure histories, but again there was no significant association (P =0.104). There was no significant difference in mean PCD between patients with chronic partial seizures and those with generalised seizures (P =0.76) and in patients with chronic partial seizures either with or without secondary generalised seizures (P = 0.56). In addition there was no significant difference in the mean PCD between patients with and without hippocampal sclerosis (P =0.62).
In a single patient who had not received phenytoin treatment (case 6) only a mild and focal loss of Purkinje cells was observed in the posterior lobe with a mean PCD of 6.9/mm. In two patients with recorded episodes of phenytoin toxicity, the mean PCD were 1.5 and 1.8/mm and in both, Purkinje cell loss predominantly affected the posterior lobes.
4. Discussion In 1930 Spielmeyer documented the occurrence of cerebellar atrophy in patients with epilepsy, predating the introduction of phenytoin in 1939 and implicating seizure mediated cell loss. Recent studies in humans and animals have primarily focused on the contribution of phenytoin to cerebellar atrophy (Ballenger et al., 1982; Lindvall and Nilsson, 1984; Dam, 1987; Botez et al., 1988; Ney et al., 1994; Luef et al., 1996; Pulliainen et al., 1998). Other factors including hypoxicischaemic nerve cell injury occurring during prolonged seizures and perinatal cerebellar injury may also contribute to the atrophy creating a complex picture to unravel. In the seizure mediated model however, discharges along cerebro– cerebellar connections are implicated in the mechanism of Purkinje cell loss, and therefore the cortical localisation or epicentre of the seizures may have some influence on the pattern of cerebellar atrophy observed. Our quantitative neuropathological study aimed to address if there was any relationship between the localisation of the neocortical pathology likely to be the cause of seizures and the distribution of cerebellar atrophy. Our parameter was loss of Purkinje cells as they are vulnerable to both excitotoxic and phenytoin mediated cellular damage (Dam et al., 1984; Savic´ and Thorell, 1996; Tauer et al., 1998) and therefore likely to be a sensitive measure of the degree of cerebellar atrophy in epilepsy. Significant reductions in mean PCD were observed in patients with epilepsy compared to controls (3.2 vs. 6.4/mm, P= 0.015) and in the eight patients where the cerebellar atrophy was not grossly detectable either by neuroimaging or
Hemisphere
Right
Region
Anterior
Superior
Horizontal
Inferior
Tonsil
Anterior
Superior
Horizontal
Inferior
Tonsil
Controls n =4 (SD) Epilepsy n=16 (SD)
6.7 (0.59)
6.1 (0.38)
5.5 (0.46)
5.9 (0.41)
6.2 (0.84)
7.01 (0.48)
6.09 (0. 46)
5.6 (0.35)
5.3 (0.61)
7.0 (1.0)
4.0 (2.0)
3.4 (2.2)
2.95 (2.6)
2.62 (2.2)
3.61 (3.0)
4.1 (2.1)
3.8 (2.4)
2.4 (2.3)
3.0 (2.6)
3.4 (2.9)
P=0.018
P= 0.0001
P= 0.002
P =0.0001
P =0.007
P= 0.0001
P= 0.003
P= 0.0001
P= 0.004
P=0.01
2.9 (1.8)
3.1 (2.1)
4.5 (2.4)
4.1 (2.09)
5.4 (2.7)
3.8 (2.7)
3.6 (2.4)
3.3 (2.4)
4.5 (2.0)
5.6 (2.7)
P= 0.02
P= 0.03
P= 0.43
P =0.12
P =0.2
P= 0.05
P= 0.08
P= 0.1
P= 0.2
P=0.38
4.5 (1.8)
3.5 (2.4)
2.2 (2.4)
1.9 (1.9)
2.7 (2.8)
4.3 (2.0)
3.9 (2.5)
2.0 (2.3)
2.2 (2.5)
2.4 (2.5)
P=0.003
P= 0.005
P =0.001
P= 0.0001
P= 0.003
P =0.01
P =0.01
P = 0.0001
P = 0.003
P=0.0001
Epilepsy Gp 1 n= 5 (SD) Epilepsy Gp 2 n= 11 (SD)
Left
R. Crooks et al. / Epilepsy Research 41 (2000) 63–73
Table 3 Results of Purkinje cell densities in cerebellar regions in controls and epilepsy patients, groups 1 and 2
69
70
R. Crooks et al. / Epilepsy Research 41 (2000) 63–73
Fig. 3. (a) Macroscopic and (b) histological section stained with haematoxylin and eosin of the cerebellar hemisphere of patient 1 to illustrate gross predominant anterior lobe atrophy and similarly (c) and (d) from patient 8 to illustrate predominant posterior lobe atrophy in epilepsy (Bar = 1 cm).
R. Crooks et al. / Epilepsy Research 41 (2000) 63–73
macroscopic examination the PCD was lower (4.5/mm, P= 0.06) confirming the sensitivity of this method. Comparing PCD in hemispheric regions revealed two distinct patterns of atrophy involving predominantly either the anterior (group 1) or posterior (group 2) lobes. Cortical malformations or lesions likely to be the cause of seizures in group 1 were identified in the occipital, parietal and temporal lobes and in group 2 in the frontal, temporal, parietal and occipital lobes. Hippocampal sclerosis was also present in three patients from each group. In addition, six of the 11 patients in group 2 had old fronto-temporal contusions but none from group 1. Anatomical information regarding cerebro – cerebellar connections (cortico-ponto-cerebellar tracts) is incomplete but it is known that projections from the motor and premotor area are crossed, and mainly terminate in the anterior cerebellar lobe whereas the posterior lateral cerebellar lobes receive their main input from parietal, temporal, frontal and cingulate neocortex with projections from the mesial temporal lobe being predominantly ipsilateral (Ito, 1984; Brodal and Bjaalie, 1979; Middleton and Strick, 1998; Savic´ and Thorell, 1998; Schmahmann and Sherman, 1998). Our quantitative analysis disclosed a component of crossed cerebellar atrophy in cases with lateralising cortical pathology despite a lack of macroscopic asymmetry; more dramatic crossed atrophy occuring in epilepsy via cerebro – cerebellar pathways is well documented (Tan and Urich, 1984). Cerebro–cerebellar pathway mediated damage could also have influenced the pattern of predominant posterior lobe atrophy in group 2 based on the localisation of neocortical pathology. The additional finding of frequent frontotemporal contusions in group 2 was striking; if these post traumatic lesions are also epileptogenic, they may have influenced this pattern of cerebellar atrophy through similar pathways. However, as cerebellar scarring and Purkinje cell loss involving particularly the tonsil and adjacent posterior lobe is described in boxers brains and following head injury (Corsellis et al., 1973) an alternative explanation is that both the cerebellar atrophy and
71
contusions represent post traumatic lesions incurred during a seizure. The pattern of posterior lobe atrophy also bears similarities to the ‘post-ictal’ pattern described in Gessaga and Urich’s original study (1985). In group 1, damage mediated via cerebro–cerebellar pathways does not explain the pattern of anterior lobe atrophy observed. We note that three of the five patients in this group had severe, poorly controlled seizures (4 or more/week, cases 1, 2 and 3) compared to two of 11 patients from group 2. It is known that cerebro–cerebellar connections are bi-directional (Middleton and Strick, 1998) with the anterior lobe exerting an inhibitory effect on cortical activity. Thus although we can not explain the aetiology of anterior lobe atrophy, we can postulate that it may have a greater influence on seizure control. The documentation of poorer seizure control in the presence of cerebellar atrophy has been previously noted (Specht et al., 1997). Hypoxic-ischaemia cellular injury occurring in cerebellar vascular territories is unlikely to have influenced the two patterns of atrophy observed as less selective Purkinje cell loss and greater basket and granule cell loss would be expected. During episodes of global hypoxia occurring in prolonged seizures, the watershed zone would be expected to show greater cell loss which was not the finding from our analysis. Cerebellar atrophy has been documented following acute and chronic phenytoin administration in clinical and neuroimaging studies (Lindvall and Nilsson, 1984; Botez et al., 1988; Ney et al., 1994; Luef et al., 1996; Pulliainen et al., 1998). The mechanism of cellular toxicity of phenytoin is unknown; animal studies report alterations in Purkinje cell axon morphology (Tauer et al., 1998) and pathological reports from humans suggest it gives rise to a diffuse loss of Purkinje cells (Gessaga and Urich, 1985). The single patient in this study not treated with phenytoin showed only minor loss of Purkinje cells and in two patients with recorded episodes of phenytoin toxicity, the posterior cerebellar lobes showed greater cell loss. All other patients had received phenytoin at some time; although we lack complete information re-
72
R. Crooks et al. / Epilepsy Research 41 (2000) 63–73
garding the duration of their treatment it is unlikely that this drug alone has caused the distinct patterns of atrophy observed but possibly contributed to the overall cell loss. There was no pattern to the seizure types experienced in the two groups; both included patients with partial seizures and it has been noted from previous in-vivo studies that cerebellar atrophy can occur in patients with partial complex or focal epilepsy in the absence of generalised seizures (Ney et al., 1994; Savic´ and Thorell, 1998). We note that one patient in group 1 had febrile convulsions and that two had episodes of status epilepticus which may be of relevance to the anterior lobe atrophy observed. There was a trend for greater Purkinje cell loss with increased duration of seizures, which is supported by observations in a previous in-vivo study (Botez et al., 1988) but this may relate also to the length of exposure to phenytoin as well as cumulative seizure induced cellular damage. We also observed a trend for lower PCD with increasing age of the patient. It is known that minor reductions in Purkinje cell number can occur after the age of 60 (Hall et al., 1975) and as six of the patients in the present study were over 60 we cannot exclude that a component of age related Purkinje cell loss may have contributed to the overall cell loss in these cases. In conclusion, this is the first quantitative neuropathological study of cerebellar atrophy in patients with epilepsy. It has confirmed that cerebellar damage in terms of Purkinje cell loss may be identified in the absence of grossly visible atrophy. We have identified two patterns of hemispheric atrophy in epilepsy patients; posterior lobe atrophy may be related to episodes of head injury, wheras anterior lobe atrophy is not. Phenytoin may act in synergy with other factors but we conclude it is unlikely alone to be the cause of the cerebellar atrophy.
Acknowledgements We would like to thank Dr M. Groves for his advice with this study.
References Ballenger, C.E., Lucke, J.F., King, D.W., El Gammal, T., Brooks, B.S., Green, J.P., 1982. Cerebellar atrophy in epilepsy and headache: lack of relationship to phenytoin. Neurology 32, 910 – 912. Botez, M.I., Ezzedine, A., Vezina, J.L., 1988. Cerebellar atrophy in epileptic patients. Can. J. Neurol. Sci. 15, 299 – 303. Brodal, P., Bjaalie, J.G., 1979. Salient anatomic features of the cortico-ponto-cerebellar pathway. In: de Zeeuw, C.I., Strata, P., Voogd, J. (Eds.), Progress in Brain Research, vol. 114, pp. 227 – 249. Corsellis, J.A.N., Bruton, C.J., Freeman-Browne, D., 1973. The aftermath of boxing. Psychol. Med. 3, 270 – 303. Dam, M., 1987. Neuropathology of the cerebellum. In: Wolf, P., Dam, M., Janz, D., Dreifus, F.E. (Eds.), Advances in Epileptology, vol. 16, pp. 15 – 20. Dam, M., Bolwig, T., Hertz, M., Bajorec, J., Lomax, P., Dam, A.M., 1984. Does seizure activity produce Purkinje cell loss? Epilepsia 25, 747 – 751. Gessaga, E.C., Urich, H., 1985. The cerebellum of epileptics. Clin. Neuropathol. 4, 238 – 245. Hall, T.C., Miller, K.H., Corsellis, A.N., 1975. Variations in the human Purkinje cell population according to age and sex. Neuropathol. Appl. Neurobiol. 1, 267 – 292. Ito, M., 1984. The Cerebellum and Neural Control. Raven Press, New York, pp. 14 – 15. Lindvall, O., Nilsson, B., 1984. Cerebellar atrophy following phenytoin intoxication. Ann. Neurol. 16, 258 – 260. Luef, G., Burtscher, J., Kremser, C., Birbamer, G., Aichner, F., Bauer, G., 1996. Magnetic resonance volumetry of the cerebellum in epileptic patients after phenytoin overdosages. Eur. Neurol. 36, 273 – 277. Middleton, F.A., Strick, P.L., 1998. Cerebellar output: motor and cognitive channels. Trends Cogn. Sci. 2, 348 – 354. Ney, G.C., Lantos, G., Barr, W.B., Schaul, N., 1994. Cerebellar atrophy in patients with long term phenytoin exposure and epilepsy. Arch. Neurol. 51, 767 – 771. Pulliainen, V., Jokelainen, M., Hedman, C., Pammo, O., 1998. A case of cerebellar atrophy after phenytoin intoxication: neurologic, neuroradiologic and neuropsychological findings. J. Epilepsy 11, 241 – 247. Savic´, I., Thorell, J.O., 1996. Localised cerebellar reductions in benzodiazepine density in human partial epilepsy. Arch. Neurol. 53, 656 – 662. Schmahmann, J.D., Sherman, J.C., 1998. The cerebellar cognitive affective syndrome. Brain 121, 561 – 579. Specht, U., May, T., Rohde, M., Ebner, A., Schmidt, R.C., Schutz, M., Wolf, P., 1997. Cerebellar atrophy and prognosis after temporal lobe resection. J. Neurol. Neurosurg. Psychiatry 62, 501 – 506. Spielmeyer, W., 1930. The anatomic substratum of the convulsive state. Arch. Neurol. Psychiatry 23, 869 – 875.
R. Crooks et al. / Epilepsy Research 41 (2000) 63–73 Tan, N., Urich, H., 1984. Post-ictal cerebral hemiatrophy: with a contribution to the problem of crossed cerebellar atrophy. Acta Neuropathol. 62, 332–339.
.
73
Tauer, U., Knoth, R., Volk, B., 1998. Phenytoin alters Purkinje cell axon morphology and targeting in vitro. Acta Neuropathol. 95, 583 – 591.
Patterns of cerebellar atrophy in patients with chronic epilepsy: a quantitative neuropathological study R. Crooks a, T. Mitchell b, M. Thom a,* a
Department of Neuropathology, Institute of Neurology, Uni6ersity College London, Queen Square, London WC1N 3BG, UK b Department of Neurology, Institute of Neurology, Uni6ersity College London, Queen Square, London WC1N 3BG, UK Received 17 August 1999; received in revised form 7 April 2000; accepted 11 April 2000
Abstract Cerebellar atrophy occuring in patients with chronic epilepsy is considered either to be a sequel of cumulative seizure-mediated cell loss or a side effect of phenytoin treatment but there is little neuropathological data regarding the distribution of this cerebellar damage. We aimed to address if there is any relationship between the localisation of the cortical pathology in symptomatic epilepsy and the pattern of neocerebellar atrophy. A quantitative neuropathological post mortem analysis of the lobular distribution of hemispheric cerebellar atrophy in 16 patients with chronic epilepsy and four controls was carried out. Cerebellar atrophy, as measured by significant reductions in hemispheric linear Purkinje cell densities was confirmed in the epilepsy patients (P= 0.015) and even where the cerebellum appeared macroscopically normal, Purkinje cell loss was evident (P = 0.062). Two distinct patterns of atrophy were observed, predominantly involving either the anterior or posterior cerebellar lobes. Posterior lobe atrophy was more often associated with old fronto-temporal contusions and may be post traumatic in aetiology rather than a result of excitotoxic damage mediated via cerebro – cerebellar pathways. As the majority of patients showing either pattern of atrophy had received phenytoin treatment, we concluded that it is unlikely that this drug acts alone in inducing the Purkinje cell loss. © 2000 Elsevier Science B.V. All rights reserved. Keywords: Cerebellar atrophy; Purkinje cell density; Epilepsy
1. Introduction Cerebellar atrophy has long been recognised from both neuropathological and neuroimaging studies to occur in a proportion of patients with chronic epilepsy (Spielmeyer 1930; Dam 1987; Botez et al., 1988). This atrophy has been variably * Corresponding author. Fax: +44-171-9169546. E-mail address: [email protected] (M. Thom)
attributed to seizure mediated cellular damage, a toxic side effect of phenytoin, transneuronal cerebellar degeneration following crossed cerebral hemiatrophy, anoxic-ischaemic injury during seizures or a combination of these factors acting synergistically (Ney et al., 1994; Savic´ and Thorell, 1996; Specht et al., 1997). Despite its long recognition there have been very few pathological studies regarding the anatomical distribution of cerebellar atrophy in epilepsy. The most detailed
0920-1211/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 0 - 1 2 1 1 ( 0 0 ) 0 0 1 3 3 - 9
64
R. Crooks et al. / Epilepsy Research 41 (2000) 63–73
study to date was based on a description of five cases illustrating the atrophy observed with each aforementioned processes (Gessaga and Urich, 1985); for example a ‘post-ictal pattern’ showed atrophy at the level of the horizontal fissure in the vascular watershed territory of the lateral lobes whereas phenytoin toxicity resulted in more diffuse cerebellar atrophy. There is evidence to suggest that cerebellar atrophy may be of clinical importance in patients with epilepsy. It has been associated with a worse prognosis, in terms of seizure control, following temporal lobectomy (Specht et al., 1997). Furthermore, pathology limited to the cerebellum has been associated with a cognitive affective syndrome (Schmahmann and Sherman, 1998) which may also be of relevance in patients with epilepsy (Specht et al., 1997; Pulliainen et al., 1998). In view of the above, we carried out a quantitative neuropathological analysis of the patterns of cerebellar hemispheric atrophy in post mortem material from patients with chronic epilepsy. We aimed to address if there is any relationship between the pattern of cerebellar atrophy and the identified neocortical pathology implicated in the seizure genesis to suggest seizure-mediated cell loss via cerebro–cerebellar pathways.
2. Methods Post mortem cerebellar tissue from 16 patients with epilepsy (age range 21 – 84, mean 49 years) was included in this study. All the post mortems or routine neuropathological examinations had been carried out between the years 1992 and 1998 in the Department of Neuropathology at the National Hospital for Neurology and Neurosurgery, London. The cases were selected to include a range of localised cerebral pathologies relevant to the patients epilepsy disorder (Table 1); these included hippocampal sclerosis (cases 2, 4, 5, 12, 14 and 15) which was bilateral and asymmetrical in two (cases 2 and 4) and associated with cortical hemi-atrophy in two (cases 5 and 15), end folium hippocampal sclerosis (cases 7, 11 and 13), malformations of cortical development including microdysgenesis (cases 1 and 5), grey matter
heterotopias (cases 7 and 14), vascular malformations (cases 3 and 9), glio-neuronal hamartoma (case 4) and early cerebral hypoxic-ischaemic injury with cortical ulegyria (case 11). In addition, four of the older patients had established focal cerebral and lacunar infarcts (cases 4, 7, 10 and 14) and old fronto-temporal contusions from previous head injuries, were identified in six cases (cases 7, 8, 9, 10, 14 and 16). In eight of the 16 cases no cerebellar atrophy was evident on routine macroscopic examination (cases 2, 3, 6, 7, 11, 12, 13 and 14) and this, in most cases where available, correlated with pre-mortem neuroimaging findings (Table 1). The hospital notes from these patients were also reviewed and the data are presented in Table 2. Eleven patients had complex partial seizures, nine with secondary generalised seizures and five patients had generalised tonic clonic seizures. In five patients the seizure frequency was severe, with maximum seizure frequencies of four per week (cases 1, 2, 3, 11 and 14). The EEG data obtained from review of the clinical records showed distinct cortical lateralising foci in seven patients (cases 2, 3, 5, 6, 8, 10 and 11). Despite our attempts to include patients who had not been treated with phenytoin, we were able to obtain neuropathological tissue from only one such patient (case 6). Two patients (cases 9 and 14) had recorded episodes of phenytoin toxicity with an acute reversible cerebellar syndrome. None of the patients had a documented history of alcohol or drug abuse. Similar tissue sections were sampled from the remaining cerebellar tissue of each case; a traditional plane of cut was made at right angles to the folia through the dentate nucleus to include the entire dorsoventral extent of both left and right hemispheres. An ‘anterior’ section corresponded to the anterior lobe, and ‘superior’, ‘horizontal’, ‘inferior’ and ‘tonsil’ sections represented the posterior lobe (Fig. 1). The ‘horizontal’ section included both superior and inferior semilunar lobules and the horizontal fissure and thus represented the vascular watershed area. Twenty micron thick sections were cut and stained with 0.01% cresyl violet from which the total length of the Purkinje cell layer was measured using a Leica
Table 1 Neuropathological findings in the 16 epilepsy patientsa Case
Cerebellum: macroscopic description, fixed hindbrain weight (% total brain weight), where recorded
Histological neuropathological diagnosis relevant to epilepsy
1 2 3
SUDEP SUDEP SUDEP
Atrophy of anterior lobes No atrophy No atrophy
Parietal lobe microdysgenesis Bilateral HCS L\R Right occipital lobe cortical scarring
4
Lobar pneumonia Bronchopneumonia SUDEP Peritonitis
Microcephaly (995) Left hippocampal atrophy (1390) Right occipital cavernoma (operated) (1380) Right occipital infarct (880)
Symmetrical atrophy
Left hippocampal and cortical hemiatrophy (1000) Normal (1160) Right parietal laminar heterotopias, recent left frontal infarct; old fronto temporal contusions (1465) Old frontal contusions (1220) Right frontal contusions (1100) Old fronto-temporal contusions, lacunar infarcts (1310) Left frontal and occipital ulegyria (1030) Right temporal lobe atrophy and arachnoid cyst (950) Left hippocampal atrophy (1650)
Symmetrical atrophy 90 g (9%) No atrophy No atrophy
Left occipital neuronal — glial hamartoma Bilateral HCS R\L Left temporal cortical microdysgensis; left HCS and left cortical hemiatrophy Acute neuronal changes in hippocampus Mild EFS
Atrophic 95 g (7.7%) Atrophy 108 g (9.8%) Atrophy in watershed area 120 g (9.1%)
No specific diagnosis Left amygdala telangiectasia No specific diagnosis
No atrophy No atrophy 115 g (12%)
Mild EFS Right HCS and temporal lobe ulegyria
No atrophy
5 6 7
8 9 10
SUDEP SUDEP Ca lung
11 12
SUDEP SUDEP
13
SUDEP
14
Bronchopneumonia
15
Bronchopneumonia Bronchopneumonia
16
a
Bilateral subependymal heterotopias; old fronto temporal contusions, old left PCA infarct (990) Left cerebral hemiatrophy (1180)
No atrophy 145 g (14%)
Mild left EFS; acute neuronal changes in subiculum Right HCS
Mild atrophy 120 g (10%)
Left HCS and left cortical hemiatrophy
Old contusions (975)
Atrophy, 100 g (10.2%)
No HCS
R. Crooks et al. / Epilepsy Research 41 (2000) 63–73
Cause of death Macroscopic appearance of brain (fixed brain weight (g))
SUDEP, sudden and unexpected death in epilepsy; PCA, posterior cerebral artery; HCS, hippocampal sclerosis; EFS, end folium sclerosis.
65
Case
Seizure history: duration, frequency (max.), type
1
21 (F)
20 years, 6-7/day; FC, CPS, SGS, MJ
2
21 (F)
3
23 (F)
4
76 (F)
5
59 (F)
6 7
27 (F) 70 (F)
8
44 (M)
9
47 (M)
10
69 (M)
11
40 (F)
12
67 (F)
13
27 (F)
14
49 (F)
15 16
73 (F) 84 (M)
EEG
Abnormal irregular slow waves and bilateral sharp waves and spikes. 3 years, 3–4/week; CPS, Generalized slow and SGS, SE sharp waves over left hemipshere 15 years, 2–3/day; MJ, Frequent paroxysmal CPS, SGS epileptiform discharges over right hemipshere 76 years, 7/year; CPS, SGS Irregular slow waves over both temporal lobes 46 years, 1/month; SE, Left sided focus CPS, SGS 12 years, 1/week; GTCS Left fronto-temporal focus 51 years, 1/year; CPS, SGS Irregular slow waves left temporal; right temporal spikes 37 years, 2–3/year; CPS, Right sided focus SGS 45 years, ‘frequent’; GTCS Abnormal irregular slow waves; bilateral sharp waves and spikes 62 years; CPS Right temporo-parietal focus 32 years, 18/month; CPS, Left temporal focus SGS 64 years, 8–12/month; Bilateral central slow wave GTCS disturbance 16 years, 1/week; MJ, Abnormal activity both GTCS, PS L&R fronto-temporal lobes 20 years, 6–7/day; FC, Abnormal irregular slow CPS, SGS, MJ waves and bilateral sharp waves and spikes. 72 years, 1–2 month; CPS Within normal limits 68 years, 1/month; GTCS Generalized abnormalities
Neuroimaging
AED treatments
Duration (maximal dose) of Phenytoin treatment when documented
CT normal
PT, CBZ, PhB, SVP, Cl, Lam
B5 years (5–100 mg)
MRI – left HCS
PT, CBZ, Vig, Cl, SVP, 10 months (600 mg) Lam
MRI – atrophy right occipital lobe
PT, PhB, Cl, Lam, Vig, NA CBZ, SVP
CT – cerebral and cerebellar atrophy CT – left hemiatrophy and cerebellar atrophy Not done MRI – parietal lobe infarct & old frontal contusions. CT – cerebral and cerebellar atrophy. MRI – minor cerebellar atrophy
PT, PhB
B5 years
PT, SVP, CBZ, PhB, Cl CBZ, SVP PT, CBZ, PhB, Cl
B10 years No PHT NA
PT, SVP, PhB, CBZ, 23 years (150 mg) Cl PT, CBZ, PhB, ViG, Cl 2 years
CT – normal
PT, CBZ, PhB, Lam
B5 years
CT – left hemiatrophy
19 years (300 mg)
Not done
PT, CBZ, PhB, Vig, Lam, Acet PT, PhB
NA
CT-atrophy
PT, PhB, CBZ, SVP
11 months (300 mg)
CT-Normal
PT, CBZ, Lam, Cl, Vig 9 years
Not done CT – cerebral and cerebellar atrophy
PT, CBZ, Cl, PhB PT, PhB
NA 20 years+
a FC, febrile convulsion; CPS, complex partial seizures; MJ, myoclonic jerks; GTCS, generalised tonic clonic seizures; SGS, secondary generalised seizures; HCS, hippocampal sclerosis; PT, phenytoin; CBZ, carbamazepine; PhB, phenobarbitone; SVP, sodium valproate; Cl, Clonazepam; Lam, Lamotrigine; Vig, Vigabatrine; Acet, acetazolamide; NA, information not available.
R. Crooks et al. / Epilepsy Research 41 (2000) 63–73
Age (sex)
66
Table 2 Clinical data of the 16 patients with chronic epilepsya
R. Crooks et al. / Epilepsy Research 41 (2000) 63–73
Q500 image analyzer (Leica Imaging Systems, Cambridge) by a single observer (RC). The total number of Purkinje cells (defined for our study as large nucleolated neurones lying at the junctions of the molecular and granule cell layers with prominent apical dendrites) were counted at a magnification of × 20 on a Leica DM RB microscope by two independent observers (RC and MT) with good reproducibility. Linear Purkinje cell densities/mm (PCD) were then calculated for each region of each hemisphere. Mean PCD values between anatomical regions and patient groups were compared using an independent Ttest and Pearson’s correlation coefficient was used to assess for any linear associations. Identical examination and quantitative analysis was carried out on four neurologically normal controls of ages 27, 47, 52 and 56 years (mean age 45 years). The cause of death in the control cases included a road traffic accident, aplastic anaemia, bronchop-
67
Fig. 2. Histogram of the mean left and right cerebellar hemispheric Purkinje cell densities in each subregion in controls, all epilepsy patients and in the two epilepsy groups; group 1 (cases 1 – 5) showing predominant anterior lobe atrophy and group 2 (cases 6 – 16) posterior lobe atrophy.
neumonia with pulmonary fibrosis and bronchial carcinoma respectively. In none of these cases was significant neuropathological or neurodegenerative disease identified.
3. Results
Fig. 1. Figure illustrating the five areas from each cerebellar hemisphere at the level of the dentate nucleus referred to in the text & tables. (1) ‘Anterior section’ comprises the central and anterior quadrangular lobule (Anterior lobe of the cerebellum). (2) ‘Superior section’ comprises the posterior quadrangular lobule and the anterior half of the superior semilunar lobule and includes the posterior superior fissure. (3) ‘Horizontal section’ comprises the posterior part of the superior semilunar lobule and the entire inferior semilunar lobule and includes the horizontal fissure. (4) ‘Inferior section’ comprises the gracile and biventer lobules and the prepyramidal fissure. (5) ‘Tonsil section’ and comprises the cerebellar tonsil. (2–5, Posterior lobe of cerebellum).
Histological examination revealed various degrees of cerebellar atrophy with Purkinje cell loss, preservation of basket cells, Bergmann gliosis and focal granule cell depletion in all the epilepsy patients. In some cases torpedo like axon swellings were present on residual Purkinje cells as evidence of ongoing cell loss. Quantitative analysis confirmed this finding with significant reductions in Purkinje cell linear densities (PCD) in the epilepsy group 3.2/mm (SD 2.1) compared to controls 6.14/mm (SD 0.19, P= 0.015). In eight patients without macroscopically apparent cerebellar atrophy, lower mean PCD were also observed (4.5/mm, P= 0.062). In the control group, slightly higher PCD were recorded in the anterior, superior and tonsil sections than in the horizontal
68
R. Crooks et al. / Epilepsy Research 41 (2000) 63–73
and inferior sections (Table 3; Fig. 2) which reflects the normal anatomical variation in Purkinje cell densities (Hall et al., 1975). Two distinct patterns of cerebellar atrophy were observed in the epilepsy patients. The first, (group 1, cases 1–5) showed a more significant reduction in PCD compared to controls in the anterior and superior sections (Table 3, Fig. 2). A second pattern (group 2, cases 6 – 16) showed greater reduction in PCD in the horizontal, inferior and tonsil sections. In cases with more severe atrophy, these two distinct patterns were obvious from macroscopic examination (Fig. 3). In two patients (cases 5 and 15) cerebral hemiatrophy with hippocampal sclerosis was present and although crossed cerebellar atrophy was not obvious on macroscopic or histological examination alone, quantitation revealed a greater reduction in PCD in the contralateral than ipsilateral cerebellar hemisphere in both cases. In six patients (cases 2, 3, 6, 8, 10 and 11), a well defined lateralising EEG focus was identified (Table 2) and in three of these (cases 2, 3 and 11) this correlated with an underlying unilateral neocortical pathology. In two patients (cases 4 and 12) a unilateral cortical pathology or malformation was identified and considered likely to be the cause of the patients seizures. In these ten patients with lateralising cortical lesions, seven (cases 2, 4, 5, 8, 10, 12 and 15) showed reduced mean PCD in the contralateral than ipsilateral cerebellar hemisphere, but the difference was not significant (P = 0.13). A trend of lower mean PCD with increasing age of the patient was observed but there was no significant correlation (P =0.107). Also we observed a trend of lower mean PCD in patients with longer seizure histories, but again there was no significant association (P =0.104). There was no significant difference in mean PCD between patients with chronic partial seizures and those with generalised seizures (P =0.76) and in patients with chronic partial seizures either with or without secondary generalised seizures (P = 0.56). In addition there was no significant difference in the mean PCD between patients with and without hippocampal sclerosis (P =0.62).
In a single patient who had not received phenytoin treatment (case 6) only a mild and focal loss of Purkinje cells was observed in the posterior lobe with a mean PCD of 6.9/mm. In two patients with recorded episodes of phenytoin toxicity, the mean PCD were 1.5 and 1.8/mm and in both, Purkinje cell loss predominantly affected the posterior lobes.
4. Discussion In 1930 Spielmeyer documented the occurrence of cerebellar atrophy in patients with epilepsy, predating the introduction of phenytoin in 1939 and implicating seizure mediated cell loss. Recent studies in humans and animals have primarily focused on the contribution of phenytoin to cerebellar atrophy (Ballenger et al., 1982; Lindvall and Nilsson, 1984; Dam, 1987; Botez et al., 1988; Ney et al., 1994; Luef et al., 1996; Pulliainen et al., 1998). Other factors including hypoxicischaemic nerve cell injury occurring during prolonged seizures and perinatal cerebellar injury may also contribute to the atrophy creating a complex picture to unravel. In the seizure mediated model however, discharges along cerebro– cerebellar connections are implicated in the mechanism of Purkinje cell loss, and therefore the cortical localisation or epicentre of the seizures may have some influence on the pattern of cerebellar atrophy observed. Our quantitative neuropathological study aimed to address if there was any relationship between the localisation of the neocortical pathology likely to be the cause of seizures and the distribution of cerebellar atrophy. Our parameter was loss of Purkinje cells as they are vulnerable to both excitotoxic and phenytoin mediated cellular damage (Dam et al., 1984; Savic´ and Thorell, 1996; Tauer et al., 1998) and therefore likely to be a sensitive measure of the degree of cerebellar atrophy in epilepsy. Significant reductions in mean PCD were observed in patients with epilepsy compared to controls (3.2 vs. 6.4/mm, P= 0.015) and in the eight patients where the cerebellar atrophy was not grossly detectable either by neuroimaging or
Hemisphere
Right
Region
Anterior
Superior
Horizontal
Inferior
Tonsil
Anterior
Superior
Horizontal
Inferior
Tonsil
Controls n =4 (SD) Epilepsy n=16 (SD)
6.7 (0.59)
6.1 (0.38)
5.5 (0.46)
5.9 (0.41)
6.2 (0.84)
7.01 (0.48)
6.09 (0. 46)
5.6 (0.35)
5.3 (0.61)
7.0 (1.0)
4.0 (2.0)
3.4 (2.2)
2.95 (2.6)
2.62 (2.2)
3.61 (3.0)
4.1 (2.1)
3.8 (2.4)
2.4 (2.3)
3.0 (2.6)
3.4 (2.9)
P=0.018
P= 0.0001
P= 0.002
P =0.0001
P =0.007
P= 0.0001
P= 0.003
P= 0.0001
P= 0.004
P=0.01
2.9 (1.8)
3.1 (2.1)
4.5 (2.4)
4.1 (2.09)
5.4 (2.7)
3.8 (2.7)
3.6 (2.4)
3.3 (2.4)
4.5 (2.0)
5.6 (2.7)
P= 0.02
P= 0.03
P= 0.43
P =0.12
P =0.2
P= 0.05
P= 0.08
P= 0.1
P= 0.2
P=0.38
4.5 (1.8)
3.5 (2.4)
2.2 (2.4)
1.9 (1.9)
2.7 (2.8)
4.3 (2.0)
3.9 (2.5)
2.0 (2.3)
2.2 (2.5)
2.4 (2.5)
P=0.003
P= 0.005
P =0.001
P= 0.0001
P= 0.003
P =0.01
P =0.01
P = 0.0001
P = 0.003
P=0.0001
Epilepsy Gp 1 n= 5 (SD) Epilepsy Gp 2 n= 11 (SD)
Left
R. Crooks et al. / Epilepsy Research 41 (2000) 63–73
Table 3 Results of Purkinje cell densities in cerebellar regions in controls and epilepsy patients, groups 1 and 2
69
70
R. Crooks et al. / Epilepsy Research 41 (2000) 63–73
Fig. 3. (a) Macroscopic and (b) histological section stained with haematoxylin and eosin of the cerebellar hemisphere of patient 1 to illustrate gross predominant anterior lobe atrophy and similarly (c) and (d) from patient 8 to illustrate predominant posterior lobe atrophy in epilepsy (Bar = 1 cm).
R. Crooks et al. / Epilepsy Research 41 (2000) 63–73
macroscopic examination the PCD was lower (4.5/mm, P= 0.06) confirming the sensitivity of this method. Comparing PCD in hemispheric regions revealed two distinct patterns of atrophy involving predominantly either the anterior (group 1) or posterior (group 2) lobes. Cortical malformations or lesions likely to be the cause of seizures in group 1 were identified in the occipital, parietal and temporal lobes and in group 2 in the frontal, temporal, parietal and occipital lobes. Hippocampal sclerosis was also present in three patients from each group. In addition, six of the 11 patients in group 2 had old fronto-temporal contusions but none from group 1. Anatomical information regarding cerebro – cerebellar connections (cortico-ponto-cerebellar tracts) is incomplete but it is known that projections from the motor and premotor area are crossed, and mainly terminate in the anterior cerebellar lobe whereas the posterior lateral cerebellar lobes receive their main input from parietal, temporal, frontal and cingulate neocortex with projections from the mesial temporal lobe being predominantly ipsilateral (Ito, 1984; Brodal and Bjaalie, 1979; Middleton and Strick, 1998; Savic´ and Thorell, 1998; Schmahmann and Sherman, 1998). Our quantitative analysis disclosed a component of crossed cerebellar atrophy in cases with lateralising cortical pathology despite a lack of macroscopic asymmetry; more dramatic crossed atrophy occuring in epilepsy via cerebro – cerebellar pathways is well documented (Tan and Urich, 1984). Cerebro–cerebellar pathway mediated damage could also have influenced the pattern of predominant posterior lobe atrophy in group 2 based on the localisation of neocortical pathology. The additional finding of frequent frontotemporal contusions in group 2 was striking; if these post traumatic lesions are also epileptogenic, they may have influenced this pattern of cerebellar atrophy through similar pathways. However, as cerebellar scarring and Purkinje cell loss involving particularly the tonsil and adjacent posterior lobe is described in boxers brains and following head injury (Corsellis et al., 1973) an alternative explanation is that both the cerebellar atrophy and
71
contusions represent post traumatic lesions incurred during a seizure. The pattern of posterior lobe atrophy also bears similarities to the ‘post-ictal’ pattern described in Gessaga and Urich’s original study (1985). In group 1, damage mediated via cerebro–cerebellar pathways does not explain the pattern of anterior lobe atrophy observed. We note that three of the five patients in this group had severe, poorly controlled seizures (4 or more/week, cases 1, 2 and 3) compared to two of 11 patients from group 2. It is known that cerebro–cerebellar connections are bi-directional (Middleton and Strick, 1998) with the anterior lobe exerting an inhibitory effect on cortical activity. Thus although we can not explain the aetiology of anterior lobe atrophy, we can postulate that it may have a greater influence on seizure control. The documentation of poorer seizure control in the presence of cerebellar atrophy has been previously noted (Specht et al., 1997). Hypoxic-ischaemia cellular injury occurring in cerebellar vascular territories is unlikely to have influenced the two patterns of atrophy observed as less selective Purkinje cell loss and greater basket and granule cell loss would be expected. During episodes of global hypoxia occurring in prolonged seizures, the watershed zone would be expected to show greater cell loss which was not the finding from our analysis. Cerebellar atrophy has been documented following acute and chronic phenytoin administration in clinical and neuroimaging studies (Lindvall and Nilsson, 1984; Botez et al., 1988; Ney et al., 1994; Luef et al., 1996; Pulliainen et al., 1998). The mechanism of cellular toxicity of phenytoin is unknown; animal studies report alterations in Purkinje cell axon morphology (Tauer et al., 1998) and pathological reports from humans suggest it gives rise to a diffuse loss of Purkinje cells (Gessaga and Urich, 1985). The single patient in this study not treated with phenytoin showed only minor loss of Purkinje cells and in two patients with recorded episodes of phenytoin toxicity, the posterior cerebellar lobes showed greater cell loss. All other patients had received phenytoin at some time; although we lack complete information re-
72
R. Crooks et al. / Epilepsy Research 41 (2000) 63–73
garding the duration of their treatment it is unlikely that this drug alone has caused the distinct patterns of atrophy observed but possibly contributed to the overall cell loss. There was no pattern to the seizure types experienced in the two groups; both included patients with partial seizures and it has been noted from previous in-vivo studies that cerebellar atrophy can occur in patients with partial complex or focal epilepsy in the absence of generalised seizures (Ney et al., 1994; Savic´ and Thorell, 1998). We note that one patient in group 1 had febrile convulsions and that two had episodes of status epilepticus which may be of relevance to the anterior lobe atrophy observed. There was a trend for greater Purkinje cell loss with increased duration of seizures, which is supported by observations in a previous in-vivo study (Botez et al., 1988) but this may relate also to the length of exposure to phenytoin as well as cumulative seizure induced cellular damage. We also observed a trend for lower PCD with increasing age of the patient. It is known that minor reductions in Purkinje cell number can occur after the age of 60 (Hall et al., 1975) and as six of the patients in the present study were over 60 we cannot exclude that a component of age related Purkinje cell loss may have contributed to the overall cell loss in these cases. In conclusion, this is the first quantitative neuropathological study of cerebellar atrophy in patients with epilepsy. It has confirmed that cerebellar damage in terms of Purkinje cell loss may be identified in the absence of grossly visible atrophy. We have identified two patterns of hemispheric atrophy in epilepsy patients; posterior lobe atrophy may be related to episodes of head injury, wheras anterior lobe atrophy is not. Phenytoin may act in synergy with other factors but we conclude it is unlikely alone to be the cause of the cerebellar atrophy.
Acknowledgements We would like to thank Dr M. Groves for his advice with this study.
References Ballenger, C.E., Lucke, J.F., King, D.W., El Gammal, T., Brooks, B.S., Green, J.P., 1982. Cerebellar atrophy in epilepsy and headache: lack of relationship to phenytoin. Neurology 32, 910 – 912. Botez, M.I., Ezzedine, A., Vezina, J.L., 1988. Cerebellar atrophy in epileptic patients. Can. J. Neurol. Sci. 15, 299 – 303. Brodal, P., Bjaalie, J.G., 1979. Salient anatomic features of the cortico-ponto-cerebellar pathway. In: de Zeeuw, C.I., Strata, P., Voogd, J. (Eds.), Progress in Brain Research, vol. 114, pp. 227 – 249. Corsellis, J.A.N., Bruton, C.J., Freeman-Browne, D., 1973. The aftermath of boxing. Psychol. Med. 3, 270 – 303. Dam, M., 1987. Neuropathology of the cerebellum. In: Wolf, P., Dam, M., Janz, D., Dreifus, F.E. (Eds.), Advances in Epileptology, vol. 16, pp. 15 – 20. Dam, M., Bolwig, T., Hertz, M., Bajorec, J., Lomax, P., Dam, A.M., 1984. Does seizure activity produce Purkinje cell loss? Epilepsia 25, 747 – 751. Gessaga, E.C., Urich, H., 1985. The cerebellum of epileptics. Clin. Neuropathol. 4, 238 – 245. Hall, T.C., Miller, K.H., Corsellis, A.N., 1975. Variations in the human Purkinje cell population according to age and sex. Neuropathol. Appl. Neurobiol. 1, 267 – 292. Ito, M., 1984. The Cerebellum and Neural Control. Raven Press, New York, pp. 14 – 15. Lindvall, O., Nilsson, B., 1984. Cerebellar atrophy following phenytoin intoxication. Ann. Neurol. 16, 258 – 260. Luef, G., Burtscher, J., Kremser, C., Birbamer, G., Aichner, F., Bauer, G., 1996. Magnetic resonance volumetry of the cerebellum in epileptic patients after phenytoin overdosages. Eur. Neurol. 36, 273 – 277. Middleton, F.A., Strick, P.L., 1998. Cerebellar output: motor and cognitive channels. Trends Cogn. Sci. 2, 348 – 354. Ney, G.C., Lantos, G., Barr, W.B., Schaul, N., 1994. Cerebellar atrophy in patients with long term phenytoin exposure and epilepsy. Arch. Neurol. 51, 767 – 771. Pulliainen, V., Jokelainen, M., Hedman, C., Pammo, O., 1998. A case of cerebellar atrophy after phenytoin intoxication: neurologic, neuroradiologic and neuropsychological findings. J. Epilepsy 11, 241 – 247. Savic´, I., Thorell, J.O., 1996. Localised cerebellar reductions in benzodiazepine density in human partial epilepsy. Arch. Neurol. 53, 656 – 662. Schmahmann, J.D., Sherman, J.C., 1998. The cerebellar cognitive affective syndrome. Brain 121, 561 – 579. Specht, U., May, T., Rohde, M., Ebner, A., Schmidt, R.C., Schutz, M., Wolf, P., 1997. Cerebellar atrophy and prognosis after temporal lobe resection. J. Neurol. Neurosurg. Psychiatry 62, 501 – 506. Spielmeyer, W., 1930. The anatomic substratum of the convulsive state. Arch. Neurol. Psychiatry 23, 869 – 875.
R. Crooks et al. / Epilepsy Research 41 (2000) 63–73 Tan, N., Urich, H., 1984. Post-ictal cerebral hemiatrophy: with a contribution to the problem of crossed cerebellar atrophy. Acta Neuropathol. 62, 332–339.
.
73
Tauer, U., Knoth, R., Volk, B., 1998. Phenytoin alters Purkinje cell axon morphology and targeting in vitro. Acta Neuropathol. 95, 583 – 591.