Traumatic compared to non-traumatic clinical-pathologic associations in temporal lobe epilepsy

EPILEPSY

~4~.c. ' , • j l

RESEARCH ELSEVIER

Epilepsy Research 19 (1994) 129-139

Traumatic compared to non-traumatic clinical-pathologic associations in temporal lobe epilepsy Gary W. Mathern a.,, Thomas L. Babb b,c, Barbara G. Vickrey u, Maria Melendez c, James K. Pretorius c a Divisions of Neurosurgery and Clinical Neurophysiology, Reed Neurological Research Center, UCLA Center for the Health Sciences, Los Angeles, CA 90024, USA b Department of Neurology, Reed Neurological Research Center, UCLA Center for the Health Sciences, Los Angeles, CA 90024, USA c Brain Research Institute, Reed Neurological Research Center, UCLA Center for the Health Sciences, Los Angeles, CA 90024, USA

Received 1 February 1994; revised 12 April 1994; accepted 15 April 1994

Abstract

This study determined differences in clinical-pathologic characteristics of intractable temporal lobe epilepsy (TLE) patients whose mechanism of cerebral injury and chronic seizures involved a prior history of cerebral trauma compared to those with non-traumatic initial injuries. TLE patients (n = 120) from a single epilepsy center were retrospectively and blindly catalogued into pathogenic groups (independent variables) based on if there was a significant Birth injury (n = 11) or Cerebral trauma (n = 26). These two 'trauma' categories were compared to TLE patients with non-seizure non-trauma histories (Non-Sz/Non-Trauma; n = 17), or a first Prolonged seizure (n = 66). The four groups were compared for differences in the time course of their clinical injuries and seizures, quantified hippocampal neuron counts, other temporal neocortical pathologies, and seizure outcomes (dependent variables). Between group statistically significant (at least P < 0.05) results showed: (1) In Birth injury, 33% had Ammon's Horn (AH) neuron loss under 50%, 54% had other temporal neocortical pathologies, they showed the most CA4 neuron loss, and the worse seizure outcomes. (2) Cerebral trauma were older when injured, 29% had AH loss under 50%, 50% showed other pathologies, and they had the best seizure outcomes. (3) Non-Sz/Non-Trauma showed the least AH and CA4 neuron losses, only 12% had other temporal pathologies, and they had seizure outcomes that were intermediate. (4) Prolonged seizure showed the youngest age of habitual TLE onsets, the greatest AH, CA1, and prosubiculum neuron loss, only 11% had other temporal pathologies, and their seizure outcomes were excellent. These results indicate that in intractable surgically treated TLE, a history of cerebral trauma or birth injury as the pathogenic mechanism of their seizures show different clinical-pathologic features and seizure outcomes compared to non-trauma patients. This supports the notion that in TLE there are different pathogenic mechanisms associated with different types of initial injuries and that patients will have different responses to surgical therapy. Keywords: Hippocampal sclerosis; Etiology; Complex partial seizures; Pathogenesis

* Corresponding author. Division of Neurosurgery, Reed Neurological Research Center, UCLA Center for the Health Sciences, Los Angeles, CA 90024-1769, USA. Tel.: (310) 206-8777; Fax: (310) 206-8461. 0920-1211/94/$07.00 © 1994 Elsevier Science B.V. All rights reserved SSDI 0 9 2 0 - 1 2 1 1 ( 9 4 ) 0 0 0 4 4 - W

130

G.W. Mathern et a l . / Epilepsy Research 19 (1994) 129-139

1. Introduction

Cerebral trauma has been associated with epilepsy since human antiquity [38], and the frequency of chronic seizures in cerebral trauma patients ranges from 3 to 28% [19]. The incidence of post-traumatic epilepsy has climbed during this century as countries have industrialized and the frequency of trauma in the population has increased [12]. Post-traumatic epilepsy has been classified by its time course into either early or late, and it is the later type that is often chronic, intractable, and is considered the most significant medically and economically [42]. Trauma is also one of the pathogenic etiologies associated with temporal lobe epilepsy (TLE) [11]. However, it is unclear how post-traumatic TLE compares clinically and pathologically to other types of post-traumatic seizures, or to other etiologic mechanisms of TLE. This study determined, in intractable surgically treated TLE patients, if the pathogenic mechanisms of birth or cerebral trauma showed clinical-pathologic characteristics different than those of TLE patients with non-trauma histories. We hypothesized that a history of significant birth injury or cerebral trauma would show differences in; (1) the time course of their clinical injuries and seizures, (2) quantified hippocampal neuron counts, (3) other temporal neocortical pathologies, and (4) seizure outcomes when compared to other TLE patients with a prior history of a first prolonged seizure or nonseizure/non-trauma initial mechanisms. This hypothesis was tested using data from a cohort of patients from a single epilepsy center evaluated and treated using standardized protocols.

2. Material and methods 2.1. Clinical material

Patients with intractable complex partial seizures of probable temporal lobe origin were evaluated at UCLA's Clinical Neurophysiology Program (CNP) from 1961 to 1992 using standardized protocols previously published and approved by the institution's Human Subject Protection Committee [16,17]. The

diagnostic approach presumed that temporal lobe seizures began in or adjacent to regions of cerebral damage, and those regions could be localized based on the anatomic area of epileptic excitability and functional deficits. Consideration of surgical therapy was based on localizing a single area for resection and did not require a history of cerebral trauma or early seizures for surgical consideration. Evaluation included detailed history and neurological examinations, interictal and ictal scalp/sphenoidal EEG, an extensive neuropsychological test battery, and intracarotid amobarbital injections (Wada test) for memory and speech representation. From 1961 until 1979 110 CNP patients had intracranial bilateral depth electrode EEG telemetry to confirm the seizure focus [9]. Thereafter, 86 of 188 patients (46%) were judged to have sufficient non-invasive clinical and neuroimaging information to localize the focus, the others required depth electrodes. Neuroimaging studies changed during this cohort's time course. Prior to 1975 imaging was limited to cerebral angiograms and pneumoencephalography. As CT, MRI, and [lsF]fluoro-2-deoxyglucose (FDG) PET technology became available, they were incorporated into the clinical protocols. In 259 patients the epileptogenic focus was localized to one temporal lobe and a standard en bloc anterior resection including 3 to 4 cm of the hippocampus was performed [8]. In 223 (86%) the pathologic specimens were of sufficient quality to study both the temporal neocortex and hippocampus. The lateral neocortex was serially sectioned and microscopically examined by the neuropathologist. For the goals of this study, 48 patients with macroscopic mass lesions [40] by pathologic examination or neuroimaging were excluded, for a final sample of 175 cases for analysis. 2.2. Clinical data collection

The purpose of this study was to discern the clinical-pathologic differences between non-lesion TLE patients with a history of birth injury or cerebral trauma compared to those with a history of a first prolonged seizure or non-seizure/non-trauma (Non-Sz/Non-Trauma) initial injuries. Thus, the important independent variables were the classification of patients into categories based on the severity and

131

G.W. Mathern et al. / Epilepsy Research 19 (1994) 129-139

pathogenic mechanisms of their initial injuries and not necessarily on any prior clinical label. The following method was used. Information concerning the presence or absence of any significant medical event prior to the onset of their habitual TLE was abstracted from the medical record. These data were collected in a uniform manner by one author (GWM) who used a standardized format for data collection. Each medical record was carefully reviewed with special attention to the multiple interviews each patient had with medical personnel as part of the CNP evaluations. Any medical event or cerebral injury was noted and designated as significant if it was associated with unconsciousness for more than 30 minutes or alteration in cognition for more than four hours. For example, a note in the medical record of a prior 'head injury' was insufficient. The interviewers' notes had to contain how the injury occurred (i.e. a fall of 10 feet) and details of the patients condition after the injury (i.e. unconscious and hospitalized). Information was deemed reliable if similar data were obtained in separate notes by different interviewers. If details were lacking from the record, attempts were made to contact the patient or a family member. These data were collected without knowledge of the hippocampal pathology or the seizure outcomes and catalogued into the following pathogenic categories: a. Birth injury: Patients having histories with clear signs of a significant problem associated with delivery such as cerebral hypoxia, evidence of scalp or skull trauma, signs of neurological deficit, or developmental delay. Some ( 4 / 1 1 ) of these patients also had seizures during the neonatal and childhood period. However, at some time point the clinical character of the seizures changed into the habitual TLE for which they were referred for treatment. b. Cerebral trauma: These events were clearly associated with significant cerebral injury with hospitalization, alterations in cognition, or coma. Examples included falls down stairs or motor vehicle accidents. To be recorded in this group the cerebral trauma event did not occur during a concomitant temporal lobe seizure but must have preceded by several months or years the onset of habitual TLE. c. N o n - S z / N o n - T r a u m a : This category included

Table 1 Number and percent of temporal lobe epilepsy patients with injuries as defined in this study from 1961 to 1992 Number Percent of non- Percent of lesion TLE all TLE A: Birth injury B: Trauma C: Non-Sz/Non-Trauma a D: Prolonged Sz Total

11 26 17 66 120

6 15 10 38 69

5 12 8 30 55

a Patients with initial injuries that were non-traumatic and nonseizure related. Cases included five with hypoxia; five with pneumonia requiring a ventilator; four with encephalitis/meningitis; two with nephritis and coma; and one with Reye's syndrome.

events that were of medical significance but were not trauma in origin and did not have a history of seizure activity noted at the time of the injury. Several different clinical diagnoses fit into this category, and patients had histories of prolonged hospitalization or required at least 12 to 24 hours to recover neurologic function (see Table 1). It is important to emphasize that these patients may have had subclinical seizures at the time of their injury, but neither the clinical record nor information from the family indicated any observable or sustained motor seizures. d. Prolonged seizure: This group had a significant medical illness and a motor seizure. A first Prolonged seizure was defined as greater than 30 minutes or a complex seizure without neurologic recovery between the individual seizures. The distinction from the N o n - S z / N o n - T r a u m a group was the presence of the motor seizure. Patients with a clinical diagnosis such as meningitis could be classified into either group. Some patients ( 1 8 / 6 0 ) had a diagnosis of childhood epilepsy after their first prolonged seizure, but there was one significant insult recalled in the history. At some point in their clinical histories the clinical character of the seizures changed into the habitual TLE for which they were referred for surgical management. The dependent variables included reliable measures about; (1) the time course of the patients' seizures, (2) quantified hippocampal neuron densities, (3) other temporal lobe neocortical pathologies, and (4) seizure control following surgery. The clini-

132

G.W. Mathern et al. / Epilepsy Research 19 (1994) 129-139

cal variables abstracted from their medical record were: (1) age at the time of their injury; (2) onset of spontaneous habitual TLE defined as that age when the patient's typical habitual seizures were first recognized by physicians a n d / o r family as distinct events; (3) age at surgery; (4) gender; and (5) side resected. For each patient, a Latent period was determined defined as the interval in years from the initial injury until the onset of their intractable habitual TLE. The Duration of Habitual TLE was the interval between the onset of TLE until surgery (see Fig. 1).

2.3. Seizure outcome data Patient follow-up was obtained using a standardized format abstracted from the medical record and research files. This information was collected independently of the above clinical information. Outcome was classified using a modification of the scale proposed by Engel [14] and was based on the incidence of seizures for the most recent 12-month period. The number of years since surgery was also recorded. Groups were catalogued as completely seizure free and/or having simple sensory seizures only (auras; I), fewer than six complex partial seizures per year (II), between 0.5 and two seizures per month (III), and more than two seizures per month (IV). The data were checked for accuracy for a subset of patients by comparison to a 1990 mailed survey that was collected and maintained separately [39].

2.4. Quantified hippocampal neuron densities As previously published [5,6,32,33] hippocampal sections were stained with cresylecht violet (CV) for histopathologic review (30 /xm thick) and for cell counts (10 /.~m). Counts were at 400 × using grid morphometric techniques with Abercrombie's correctiort for section thickness, and the hippocampal subfields counted were based on Lorente de N6s [28] classification. The subfields were the upper and lower granule cell blades of the fascia dentata, and CA4, CA3, CA2, CA1 stratum pyramidale, prosubiculum, and subiculum neurons. The counts were performed by one person (JKP), blinded to the group classifications. Human control comparison tissue consisted of autopsies of comparable ages without cerebral

pathology, processed in the same manner as the surgical material, and collected within a few hours after death prior to autolysis. Previous published studies [5,6,31,32,36] have found that autopsy controls as used in this study do not introduce artifacts that affect human neuron density measurements, and control hippocampal cell densities are stable from age 2 to 55 years. It must be emphasized that all statistical tests were performed using the neuron density measurements. However as in prior publications and for easier interpretation, results were presented as the percent of neuron loss relative to controls.

2.5. Other temporal lobe pathologies The pathology reports were abstracted, and any cortical or white matter abnormalities were noted. Microscopic cortical dysplasia and Chaslin's gliosis were considered insignificant, based on previous criteria and findings [4,32].

2.6. Data analysis Data were entered into a database on a personal computer and analyzed using a statistical program (Super ANOVA Version 1.1, Abacus Concepts, Inc., Berkeley CA). Differences between the injury groups were statistically compared using an analysis of variance (ANOVA) and further compared between individual groups (at P < 0.05) using the Games-Howell test that controls for multiple comparisons of unequally sized samples and of unassumed variances. Other statistical tests included chi-square. Results were plotted with DeltaGraph Professional (DeltaPoint, Inc., Monterey, CA).

3. Results

3.1. Clinical profile Of the 175 non-lesion temporal lobe/hippocampal specimens, 120 (69%) were catalogued into one of the four groups as shown in Table 1. In seven patients clinical information was not obtainable, in six cases multiple and different cerebral injuries had occurred before the onset of their habitual TLE, and

G.W. Mathern et al. /Epilepsy Research 19 (1994) 129-139 these 13 patients were excluded from the analysis. The other 42 non-lesion cases had either repetitive minor seizures or idiopathic TLE and have been presented elsewhere [34]. Of the patients with 'trauma' (A and B) a seizure was noted with the initial event in one case of Cerebral trauma (B). There were 54 left-sided resections (45%; P = NS), and there was no statistical differences within or between the catalogued groups. No gender differences between groups were found. In the 'trauma' groups (A and B) the mechanism of injury was age dependent. In Birth injury (A; 7 f e m a l e s / 4 males) there was historical evidence of hypoxia in seven cases (64%; blue inactive baby with a nuchal cord and later developmental delay, etc.) The other four had histories of direct trauma to the calvarium. In one, a forceps delivery was associated with a large scalp hematoma on the same side as the eventual epileptogenic temporal lobe. Another baby, a twin in breech position delivered vaginally, was noted to have a right-sided collapsed lung, a right-sided transient brachial plexus palsy, and the eventual seizure focus was in the right temporal lobe. The third case had a subtemporal decompression within two days of birth for presumed hydrocephalus with a complication of a CSF leak that later developed into a seizure focus on that same side. In the Trauma category (B; 14 females/12 males) injuries occurred over a wide age range (0.2 to 29 years; median 2.5 years), and involved relatively low velocity mechanisms in 20 (77%) cases (e.g. falls, dropped some distance, or blunt blows to the head). The other six cases were associated with high velocity accidents (e.g. pedestrian and passenger motor vehicle accidents). One case (4%) had an identified cerebral hematoma and another a skull fracture at the time of the injury. Using the head injury scale as described by Frankowski et al. [19] fifteen cases were classified as severe, six as moderate, five as mild, and none as trivial. The clinical diagnoses and mechanisms for the N o n - S z / N o n - T r a u m a (C; nine female/eight male) group are shown in Table 1.

3.2. Clinical features of the TLE Fig. 1 illustrates the averaged time course of the initial injuries and clinical seizures for each of the groups. By definition the age at Birth injury (A) was

MEAN IPI ONSET

133

MEAN HABITUAL SEIZURE ONSET

TRAUMA

MEAN AGE AT SURGERY

i

NON-SEIZURE NON-TRAUMA

~

~

L

SEIZURES 0 ....

5 ....

{0 . . . .

1'5 . . . .

2r0 . . . .

25 . . . .

3J0. . . .

35

YEARS

Fig. 1. Graph illustrating the clinical features of the different TLE injury groups. The mean age (years+ SEM; large closed arrows left side of graph) for Birth injury (0 by definition), Trauma (6.3 + 1.6), Non-Sz/Non-Trauma (5.5 + 1.5), and Prolonged seizures (2.0 + 0.4) were statistically different 0~r~ov~ F = 7.4; P = 0.0001), with Birth injury significantly younger (P < 0.05; Games-Howell) than the others. The age of habitual TLE onsets (open arrows) for Birth injury (11.3+2.0), Trauma (18.1 + 1.8), Non-Sz/Non-Trauma (17.2+1.9), and Prolonged seizures (11.2 +0.8) were significantly different (F = 7.4; P = 0.0001), with the Prolonged seizure group significantly (P < 0.05) younger than the Trauma and Non-Sz/Non-Trauma groups. The age at surgery (small closed arrows) ranged from 28.3+1.1 to 32.1+1.4 and was not statistically different (F = 1.3; P = 0.27). The length of the Latent period ranged from 9.3 + 0.8 to 11.8± 2.2 and was not statistically different ( F = 1.1; P = 0.34). Also, the length of habitual TLE ranged from 13.9+1.6 to 17.2-1-1.8 and was not quite statistically different (F = 2.3; P = 0.077). 0 years. However, the age (mean + SEM years) at injury for the other groups (large closed arrows) differed significantly, with the Cerebral trauma group (B) the oldest (see legend Fig. 1). The age of habitual TLE onsets for the four groups showed a similar statistically significant difference (open arrows Fig. 1), but the age at surgery was not different (small closed arrows Fig. 1). The Latent intervals and a duration of habitual TLE varied from 9 . 2 _ 0.7 to 16.8 ___1.1 years and were not statistically different within or between groups (see legend Fig. 1).

3.3. Temporal lobe pathologies and hippocampal neuron losses Fig. 2 shows the amount and pattern of neuron losses for the different injury groups. The A m m o n ' s Horn neuron loss (illustrated as the percent A m m o n ' s Horn loss; %AH) was significantly different between the four groups ( P = 0.026; Fig. 2), and the Non-

134 ~t 1

G. W. Mathern et al. / Epilepsy Research 19 (1994) 129-139 • BIRTHINJURY [ ] TRAUMA

[ ] NON-SZ/NON-TRAUMA [ ] PROLONGEDSZS

p<0.~

p<0.01

80

P < 0,04

p = 0.026 *

70

q .~ 3o

'i

'~u

FOE

CA4

CA3 CA2 CA1 . . SUBFIELDS

PRO

Stm

AH

Fig. 2. Graph of the amount and pattern of hippocampal neuron loss for the different injury groups. The percent Ammon's Horn (%AH; + SEM) loss for Birth injury (60.8 + 9.1), Trauma (60.6 + 4.5), Non-Sz/Non-Trauma (54.5+4.7), and Prolonged seizures (68.3 + 1.7) were statistically different (ANOVA; F = 3.2; P = 0.026), with the Prolonged seizures group (*) showing significantly more (P<0.05; Games-Howell) loss than the NonSz/Non-Trauma group. Within the hippocampal subfields three regions, CA4, CA1, and prosubiculum (PRO) showed statistical differences (see P values on graph). The Non-Sz/Non-Trauma group showed the greatest hippocampal variability, especially less CA4 loss, the Birth injury group showed the greatest CA4 loss, and the Prolonged seizures group showed the most CA1 and PRO

least CA4 loss, Birth injury (A) the most CA4 loss, and the Prolonged seizure (D) group showed the greatest CA1 and prosubiculum losses. Hippocampal loss under 50% and the frequency of other temporal lobe pathologies varied between groups as shown in Table 2. In the Birth injury (A), Trauma (B), and N o n - S z / N o n - T r a u m a (C) categories significantly more (17; 38%) patients showed AH loss of less than 50% compared to 7 cases (11%; P < 0.01) in the Prolonged seizure (D) group. Other neocortical temporal lobe pathologies were found significantly more frequently in the Birth injury (A) and Trauma (B) groups compared to the other two categories ( P < 0.01; Table 2). These pathologies reflected signs of old trauma such as cortical damage (n = 15) or white matter scars (n = 4) compared to the pathologies in the non-trauma groups (C and D; i.e. dermoid cyst, arachnoid cyst, and venous angioma). As might be expected, the incidence of significant hippocampal neuron loss and another pathology in the temporal neocortex was significantly more frequent ( P < 0.01) in the Birth injury (A) and Cerebral trauma (B) categories compared to the non-trauma groups (C and D; Table 2). 3.4. S e i z u r e o u t c o m e

loss.

S z / N o n - T r a u m a (C) category showed significantly ( P < 0.05) less hippocampal damage compared to the Prolonged seizure group (D). The N o n - S z / N o n Trauma (C) and Birth injury (A) groups showed the most variable neuron losses in the different subfields (Fig. 2). Three hippocampal subfields, CA4, CA1, and prosubiculum, showed statistically different neuron losses for the different injury groups (see legend Fig. 2). The N o n - S z / N o n - T r a u m a (C) showed the

Data at least one year after surgery were available for 112 (93%) patients, and the average follow-up (mean _ SEM) did not differ between the categories (Fig. 3). The relief of habitual TLE for the four categories differed significantly ( P = 0.0043), and patients with Birth injury (A) had statistically significant worse outcomes ( P < 0.05) compared to the Cerebral trauma (B) and Prolonged seizure (D) categories. Patients with N o n - S z / N o n - T r a u m a (C) showed numerically fewer seizure free (Class I) and

Table 2 The variability of hippocampal neuron loss and number of cases with other temporal lobe pathologies < 50% All loss a Other temporal lobe Temporal lobe and (percentage) pathologies b (%) hippocampal sclerosis b A: Birth injury 3/9 (33) 6/11 (54) 4 (44) B: Trauma 6/21 (29) 13/26 (50) 7 (33) C: Non-Sz/Non-Trauma 8/~5 (53) 2/16 (12) 1 (7) D: Prolongedseizures 7/61 (11) 3/66 (5) 3 (5) and C compared to D; chi-square; P < 0.01. b A and B compared to C and D; chi-square; P < 0.01. a A, B,

G.W. Mathern et al. / Epilepsy Research 19 (1994) 129-139 100-

m o r e C l a s s III p a t i e n t s t h a n e i t h e r the T r a u m a ( B ) or P r o l o n g e d s e i z u r e ( D ) g r o u p s , b u t the d i f f e r e n c e w a s

9080-

n o t statistically significant.

706050403020-

135

~ •

. • OUTCOME CLASS

~

BIRTH INJURY N= 11 AVER E/U + SEM 7.6+1.7

TRAUMA 25 7.5+1.1

NON-SZ PROLONG NON-TR SZS 16 60 4 . 9 + 1 . 1 6.3+0.67

Fig. 3. Graph of the seizure outcomes classes for the different injury groups for the one year prior to their last follow-up. Outcome class was as follows: Class I, seizure free with or without auras; II, less than six seizures per year; III, 0.5 to two seizures per month; and IV, more than two seizures per month. The differences between the four groups were statistically significant ( F = 4.6; P = 0.0043), and comparisons between individual groups showed that the Birth injury groups were significantly ( P < 0.05) worse than the Trauma and Prolonged seizure groups. Non-Sz/Non-Trauma patients showed numerically more Class II and III outcomes, but the difference was not statistically significant. The lengths of the follow-ups were not significantly different ( F = 0.98; P = 0.40).

4. Discussion T h e m a j o r f i n d i n g s are s u m m a r i z e d in T a b l e 3. In this s t u d y f r o m a s i n g l e e p i l e p s y c e n t e r o f i n t r a c t a b l e T L E patients, the d e c i s i o n for s u r g i c a l t h e r a p y w a s b a s e d o n the l o c a l i z a t i o n o f a s i n g l e area for resection a n d did n o t require a h i s t o r y o f a p r i o r i n j u r y for s u r g i c a l c o n s i d e r a t i o n . O u r results, b i a s e d t o w a r d i n t r a c t a b l e T L E patients, s h o u l d not f a v o r a n y s i n g l e c a t e g o r y or m e c h a n i s m o f initial i n j u r y for s u r g i c a l t r e a t m e n t . T h e r e l e v a n c e o f the results to t h e clinical, pathologic, and pathogenic features of traumatic c o m p a r e d to n o n - t r a u m a t i c T L E are as follows.

4.1. Birth injury and T L E I n j u r i e s at b i r t h are o n e r e c o g n i z e d s o u r c e o f c h r o n i c e p i l e p s y [21], b u t the a s s o c i a t i o n w i t h hip-

Table 3 Major findings of this study I. Statistically significant'relationships of injuries

A. Clinical variables (Fig. I) Age at injury Age at TLE onset Age at surgery Latent interval TLE duration B. Hippocampal neuron losses (Fig. 2) Percent Ammon's Horn CA4 CA1 PRO C. Outcomes (Fig. 3)

P value

Between-group differences ( P < 0.05)

P = 0.0001 P = 0.0001

Birth injury the youngest Prolonged seizures younger than Trauma and Non-Sz/Non-Trauma

P = 0.27 (NS) P = 0.34 (NS) P = 0.07 (NS) P P P P P

= < < < =

0.025 0.04 0.01 0.04 0.0043

Prolonged seizures more loss than Non-Sz/Tr

Birth injury worse than Trauma and Prolonged seizure groups

II. Other important findings (Table 2) P value A. The Prolonged seizure groups showed fewer hippocampi with < 50% Ammon's Horn loss than other groups B. Birth injury and Trauma showed more other pathologies and dual pathologies in the temporal specimen

P < 0.01 P < 0.01

136

G. W. Mathern et aL / Epilepsy Research 19 (1994) 129-139

pocampal sclerosis and TLE has been controversial. Earle et al. [13] suggested there was an important association between birth injury and TLE. They noted pathologic findings suggesting hippocampal compression or anoxia in 100 of 157 (63%) TLE cases, and proposed that the injury occurred at birth. Reports since have found a variable incidence of birth injury in TLE. Bruton [7] found a history of difficult delivery in 30 of 249 (12%) in the Maudsley series, while the Montreal group reported 28% [1]. Our US study showed an incidence of 5-6% (Table 1), and supports the notion that birth injury is probably one of the least important factors in the etiology of TLE [15]. Further, we found that the outcome in this group was significantly worse than other TLE patients (Fig. 3). This finding was not unexpected. Malamud [30] studied autopsies of adults with histories of epilepsy from birth injuries and found that 36 of 37 cases (97%) showed diffuse dorsal hemispheric damage. Further, there was a high incidence of Ammon's Horn damage (86%) which was bilateral in 75% of cases. He compared these autopsies to another adult seizure group with a history of neonatal cerebral infections and found that in contrast, the infection group showed primarily temporal lobe pathology (12 of 20; 60%), often associated with hippocampal damage (80%), and it was again often bilateral (87%). This suggests that the poorer outcomes we found in the Birth injury and NonSz/Non-Trauma groups may be due to additional and widespread cerebral damage outside or in the opposite temporal lobe than the one resected. The other damaged areas may not be as epileptogenic as the resected tissue which would explain why there was some, but limited seizure improvement following surgery. 4.2. Trauma and TLE Like birth injury, cerebral trauma has long been recognized as a possible etiology of epilepsy [12], but it is unclear what are the differences in posttraumatic TLE compared to non-traumatic TLE and to other forms of post-traumatic seizures. In surgical series of TLE the incidence of post-traumatic etiologies has been between 15% (7; present study) and 37% [11]. Our study shows that the clinical characteristics of post-traumatic TLE are different than

other post-traumatic seizures. Jennett [25] found late epilepsy which included TLE in 5% of post-traumatic cases, and late seizures were associated with three risk factors, the presence of an acute hematoma (31%), early seizures (25%), and a depressed fracture (15%). Slightly more than half of the late epilepsy began within the first year following injury, and trauma was more often a disease in males [19]. Our data show that post-traumatic TLE differs from other post-traumatic seizures in that they: (1) began on average 12 years after injury; (2) were not associated with the above risk factors; (3) were not gender dependent; and (4) the injury occurred at a younger age (average 6.3 years; median 2.5 years) than the incidence of other post-traumatic seizures [10,20]. Traumatic TLE contrasts with other pathogenic mechanisms of TLE in that: (1) the initial injury was often at an older age; (2) there was frequently hippocampal damage under 50%; (3) there was more frequent neocortical temporal pathologies consistent with old trauma (Table 2); and (4) the relief of seizures after surgery was the same as patients with an initial prolonged seizure. This later finding contrasts with Bruton's [7] conclusion that traumatic TLE had poor seizure outcomes. However, his study classified patients based on their pathologic diagnosis, not their clinical histories, and the selection criteria for surgery did not include ictal scalp or depth EEG. This would suggest that in post-traumatic TLE the pathology is often unilateral compared to the bilateral pathology in Birth injury and Non-Sz/Non-Trauma. However, the surgical specimens in patients with post-traumatic TLE often contain signs of old trauma in the temporal neocortex, and one third have hippocampal neuron counts under 50%. This contrasts with the more damage in specimens from the Prolonged seizure group (see Fig. 2 and Table 2), and may be important when considering the extent of temporal lobe removal for surgical therapy. 4.3. Pathogenesis of traumatic TLE Traumatic brain injury can occur either from the primary insult or from secondary damage such as hypoxia and increased intracranial pressure (ICP). Significant primary traumatic hippocampal injury is not a frequent finding. Courville [11] examined fatal

G. W. Mathern et al. / Epilepsy Research 19 (1994) 129-139

human closed head injuries and found contusions in the hippocampus in 11 of 108 cases (10%). Mesial temporal damage has been more frequently associated with the secondary damage in fatal cases, but these pathologic findings are usually in the parahippocampal gyrus [2,3,35]. The hippocampus in adult and pediatric fatal head injury does show qualitative damage that consists of (1) focal damage in CA1 in 73 to 84% of cases, (2) less frequent injury in the other subfields (27-53%), and (3) frequent bilateral injury [26,27]. Further, a similar pattern of hippocampal injury was found in fatal cases not associated with hypoxia or increased ICP. These human pathologic studies suggest that in fatal head injury the hippocampus can be damaged, but the amount of damage appears less than the neuron losses associated with our group of post-traumatic TLE. Further, these studies would imply that hippocampal damage is fairly frequent and does not explain the low incidence of post-traumatic TLE in patients with cerebral trauma a n d / o r post-traumatic seizures. Thus, either the incidence of hippocampal damage in non-fatal cases is less than in fatal cases, or alternately other pathologic processes in addition to the trauma are necessary to initiate the pathophysiologic process of hippocampal sclerosis and TLE. The experimental literature would suggest that severe post-traumatic hippocampal damage associated with non-fatal insults may require more than one injury. For example, experiments using the fluid-percussion model have shown that immediately following the injury there is an increase in the levels of extracellular glutamate, aspartate, and potassium that often reaches toxic levels [23,37]. This type of experimental injury is accompanied by (1) a reduction in oxidative metabolism [22], (2) accumulation of calcium at the injury site [18], and (3) an initial increase followed by a prolonged decrease in glucose utilization [41]. However, in rats the hippocampal damage beneath the cortical contusion from a single percussion injury is primarily to hilar neurons [29]. Further, greater hippocampal damage has been noted in rats exposed to a fluid percussion followed shortly by a hypoxic injury than with either insult alone [24]. This suggests that during recovery from the non-fatal traumatic insult the injured neurons are more vulnerable to a second insult and this may be one mechanism that would explain how some hippocampi in

137

head injury could sustain additional damage that may be associated with later TLE. Additional experiments will be necessary to test this hypothesis.

5. Conclusion This study has shown that in intractable TLE patients, a history of either birth injury or cerebral trauma shows different clinical, pathologic and seizure outcomes compared to patients with a history of a first prolonged seizures or non-seizure/nontrauma injury. Surgical control of chronic temporal lobe seizures was worse in the patients with a history of a birth injury and non-seizure/non-trauma injuries, which probably reflects cerebral damage that is outside the area of resection. The outcome in the traumatic-TLE patients was comparable to patients with first prolonged seizures, but the trauma groups often have additional traumatic pathology in the temporal neocortex. These results support the notion that within the clinical spectrum of TLE there are different pathogenic mechanisms associated with the type of initial injury, these mechanisms generate different clinical-pathologic characteristics, and patients will have a differing response to surgical therapy.

Acknowledgments This work was supported by NIH grant NS 02808, a Clinical Investigator Development Award to GWM (K08 NS 01603), AHCPR grant RO1 HS06856 to BGV, and the UCLA Division of Neurosurgery. The authors wish to thank the many members of the UCLA Clinical Neurophysiology Program, especially Paul Crandall, MD, who over the years have consistently collected, recorded, and maintained the patient data files and information used in this study.

References [1] Abou-Khalil,B., Andermann,E., Andermann,F., Olivier,A. and Quesney, L.F., Temporallobe epilepsy after prolonged febrile convulsions: Excellent outcome after surgical treatment, Epilepsia, 34 (1993) 878-883. [2] Adams, J.H. and Graham, D.I., The relationship between

138

[3]

[4]

[5]

[6]

[7] [8]

[9]

[10] [11]

[12]

[13]

[14]

[15] [16]

[17]

[18]

[19]

G. W. Mathern et aL / Epilepsy Research 19 (1994) 129-139 ventricular fluid pressure and the neuropathology of raised intracranial pressure, Neuropathol. Appl. Neurobiol., 2 (1976) 323-332. Adams, J.H., Graham, D.I. and Gennarelli, T.A., Contemporary neuropathological considerations regarding brain damage in head injury. In: D.P. Becker and J.T. Povlishock (Eds.), Central Nervous System Trauma Status Report-1985, NINDS, National Institutes of Health, Washington, 1985, pp. 65-77. Babb, T.L., Research on the anatomy and pathology of epileptic tissue. In: H. Liiders (Ed.), Epilepsy Surgery, Raven Press, New York, 1991, pp. 719-727. Babb, T.L., Lieb, J.P., Brown, W.J., Pretorius, J. and Crandall, P.H., Distribution of pyramidal cell density and hyperexcitability in the epileptic human hippocampus, Epilepsia, 25 (1984) 721-728. Babb, T.L., Brown, W.J., Pretorius, J., Davenport, C., Lieb, J.P. and Crandall, P.H., Temporal lobe volumetric cell densities in temporal lobe epilepsy, Epilepsia, 25 (1984) 729-740. Bruton, C.J., The Neuropathology of Temporal Lobe Epilepsy, Oxford University Press, London, 1988. Crandall, P.H., Cortical resections. In: J. Engel, Jr. (Ed.), Surgical Treatment of the Epilepsies, Raven Press, New York, 1987, pp. 377-404. Crandall, P.H., Walter, R.D. and Rand, R.W., Clinical applications of studies of stereotactically implanted electrodes in temporal-lobe epilepsy, J. Neurosurg., 20 (1963) 827-840. Cohadon, F., Richer, E. and Castel, J.P., Head injuries: Incidence and outcome, J. Neurol. Sci., 103 (1991) $27-$31. Courville, C.B., Traumatic lesions of the temporal lobe as the essential cause of psychomotor epilepsy. In: M. Baldwin and P. Bailey (Eds.), Temporal Lobe Epilepsy, Charles C. Thomas, Springfield, 1958, pp. 220-239. Dinner, D.S., Posttraumatic epilepsy. In: E. Wyllie (Ed.), The Treatment of Epilepsy: Principles and Practice, Lea and Febiger, Philadelphia, 1993, pp. 654-658. Earle, K.M., Baldwin, M. and Penfield, W., lncisural sclerosis and temporal lobe seizures produced by hippocampal herniation at birth, Arch. Neurol. Psychiatry, 69 (1953) 27-42. Engel, Jr., J., Outcome with respect to epileptic seizures. In: J. Engel, Jr. (Ed.), Surgical Treatment of the Epilepsies, Raven Press, New York, 1987, pp. 553-571. Engel, Jr., J., Seizures and Epilepsy, F.A. Davis, Philadelphia, 1989. Engel, Jr., J., Protocols for the University of California, Los Angeles. In: J. Engel, Jr. (Ed.), Surgical Treatment of the Epilepsies, 2nd edition, Raven Press, New York, 1993, pp. 743-745. Engel, Jr., J., Levesque, M., Crandall, P.H., Shewman, A., Rausch, R. and Sutherling, W., The epilepsies. In: R.G. Grossman (Ed.), Principles of Neurosurgery, Raven Press, New York, 1991, pp. 319-358. Fineman, I., Hovda, D.A., Smith, M., Yoshino, A. and Becker, D.P., Convussive brain injury is associated with a prolonged accumulation of calcium: A 45Ca autoradiographic study, Brain Res., 624 (1993) 94-102. Frankowski, R.F., Annegers, J.F. and Whitman, S., Epidemi-

[20]

[21]

[22]

[23] [24]

[25] [26]

[27]

[28]

[29]

[30]

[31]

[32]

[33]

ological and descriptive studies. Part I: The descriptive epidemiology of head trauma in the United States. In: D.P. Becker and J.T. Povlishock (Eds.), Central Nervous System Trauma Status Report-- 1985, NINDS, National Institutes of Health, Washington, 1985, pp. 33-43. Gada, G.F., Becker, D.P., Miller, J.D. and Dwan, P.S., Pathology and pathophysiology of head injury. In: J.R. Youmans (Ed.), Neurological Surgery, third edition, W.B. Saunders, Philadelphia, 1990, pp. 1965-2016. Hauser, W.A. and Kurland, L.T., The epidemiology of epilepsy in Rochester, Minnesota, 1935 through 1967, Epilepsia, 16 (1975) 1-66. Hovda, D.A., Yoshino, A., Kawamata, T., Katayama, Y. and Becker, D.P., Diffuse prolonged depression of cerebral oxidative metabolism following concussive brain injury in the rat: A cytochrome oxidase histochemistry study, Brain Res., 567 (1991) 1-10. Hovda, D.A., Becker, D.P. and Katayama, Y., Secondary injury and acidosis, J. Neurotrauma, 9 (1992) $47-$60. Jenkins, L.W., Moszynski, K., Lyeth, B.G., Lewelt, W., Dewitt, D.S., Allen, A., et al., Increased vulnerability of the mildly traumatized rat brain to cerebral ischemia: The use of controlled secondary ischemia as a research tool to identify common or different mechanisms contributing to mechanical and ischemic brain injury, Brain Res., 477 (1989) 211-224. Jennett, B., Epilepsy After Non-missile Head Injuries, William Heinemann, Chicago, 1975. Kotapka, M.J., Graham, D.I., Adams, J.H. and Gennarelli, T.A., Hippocampal pathology in fatal non-missile human head injury, Acta Neuropathol., 83 (1992) 530-534. Kotapka, M.J., Graham, D.I., Adams, J.H., Doyle, D. and Gennarilli, T.A., Hippocampal damage in fatal paediatric head injury, Neuropath. Appl. Neurobiol., 19 (1993) 128133. Lorente d6 No, R., Studies on the structure of the cerebral cortex. II. Continuation of the study of the ammonic system, J. Psychol. Neurol., 45 (1933) 113-177. Lowenstein, D.H., Thomas, M.J., Smith, D.H. and Mclntosh, T.K., Selective vulnerability of dentate hilar neurons following traumatic brain injury: A potential mechanistic link between head trauma and disorders of the hippocampus, J. Neurosci., 12 (1992) 4846-4853. Malamud, N., Some observations on the pathology of epilepsy following birth trauma and postnatal infection in the light of temporal lobe epilepsy. In: M. Baldwin and P. Bailey (Eds.), Temporal Lobe Epilepsy, Charles C. Thomas, Springfield, 1958, pp. 149-165. Mathern, G.W., Leite, J.P., Pretorius, J.K., Quinn, B., Peacock, W.J. and Babb, T.L., Children with severe epilepsy: Evidence of hippocampal neuron losses and aberrant mossy fiber sprouting during postnatal granule cell migration and differentiation, Dev. Brain Res., 78 (1994) 70-80. Mathern, G.W., Pretorius, J.K. and Babb, T.L., The quantified patterns of mossy fiber sprouting and neuron densities in hippocampal and lesional seizures, J. Neurosurg., (1994) in press. Mathern, G.W., Pretorius, J.K. and Babb, T.L., The age and type of initial injury influences the latent period, MRI hip-

G.W. Mathern et al. / Epilepsy Research 19 (1994) 129-139

[34]

[35]

[36] [37]

[38]

pocampal atrophy, Ammon's horn pathology, and surgical outcome, J. Neurosurg., (1994) in press. Mathern, G.W., Babb, T.L., Vickrey, B.G., Melendez, M. and Pretorius, J.K., The clinical-pathogenic mechanisms of hippocampal neuron loss and surgical outcomes in intractable temporal lobe epilepsy, Brain, (1994) in review. McCormick, W.F., Pathology of dosed head injury. In: R.H. Wilkins and S.S. Rengachary (Eds.), Neurosurgery, McGraw-Hill, New York, 1985, pp. 1544-1569. Mouritzen-Dam, A., Epilepsy and neuron loss in the hippocampus, Epilepsia, 21 (1980) 617-629. Palmer, A.M., Marion, D. W, Botscheller, M.L., Swedlow, P.E., Styren, S.C. and Dekosky, S.T., Traumatic brain injury-induced excitotoxicity assessed in a controlled cortical impact model, J. Neurochem., 61 (1993) 2015-2024. Temkin, O., The Falling Sickness, Johns Hopkins Press, Baltimore, 1945.

139

[39] Vickrey, B.G., Hays, R.D., Graber, J., Rausch, R., Engel, Jr., J. and Brook, R.H., A health-related quality of life instrument for patients evaluated for epilepsy surgery, Med. Care, 30 (1992) 299-319. [40] Weiser, H.G., Engel, Jr., J., Williamson, P.D., Babb, T.L. and Gloor, P., Surgically remediable temporal lobe syndromes. In: J. Engel, Jr. (Ed.), Surgical Treatment of the Epilepsies, 2nd edition, Raven Press, New York, 1993, pp. 49-63. [41] Yoshino, A., Hovda, D.A., Kawamata, T., Katayama, Y. and Becker, D.P., Dynamic changes in local cerebral glucose utilization following cerebral concussion in rats: Evidence of a hyper- and subsequent hypometabolic state, Brain Res., 561 (1991) 106-119. [42] Young, B., Post-traumatic epilepsy. In: J.R. Youmans (Ed.), Neurological Surgery, third edition, W.B. Saunders, Philadelphia, 1990, pp. 2243-2249.