Hippocampal atrophy, epilepsy duration, and febrile seizures in patients with partial seizures

Methods.

The patients (mean age 34.8 ± 10.2 years; 19 men) had been referred to the Clinical Epilepsy Section, Epilepsy Research Branch, National Institute of Neurological Disorders and Stroke, for evaluation of uncontrolled partial seizures. Patients were selected for inclusion if ictal video-EEG telemetry with sphenoidal electrodes showed CPS with a temporal lobe epileptogenic zone. Because MRI volumetric studies were the dependent variable, they were not used as an inclusion criterion. Patients with evidence of a mass lesion were excluded. Twenty-two patients had surgery, and seven had invasive electrode studies. Mean age at seizure onset was 11.4 ± 5.9 years, and epilepsy duration was 23.1 ± 11.4 years. Detailed histories were obtained from patients, their families, and medical records. All patients had a history of secondary generalization. Nine patients were considered to have had initial “prolonged or complex” FS (FS+)—a seizure associated with fever occurring before the onset of afebrile seizures and lasting longer than 15 minutes, with focal features, or followed by transient or persistent neurologic abnormalities.11 FS− patients could have a history of simple FS, as defined by Nelson and Ellenberg.11 FS+ and FS− patients did not differ in epilepsy duration, age at seizure onset, or age at scan.

Contiguous coronal MRIs with a 2-mm slice thickness were obtained on a GE Signa 1.5-T MR scanner (GE Medical Systems, Milwaukee, WI). Scanning sequence was repeat time (TR) 24 msec, echo time (TE) 5 msec with a flip angle of 45°, field of view 24 × 24 cm with a matrix of 256 × 256. The voxel size was 0.9375 × 0.9375 × 2 mm³. Images were transferred to a VAX station and analyzed using the Medical Imaging Retrieval, Analysis and Graphics software (MIRAGE, NIH, Bethesda, MD). HF, temporal lobe, and whole brain volumes were drawn on consecutive images as described previously.12 The hippocampus was outlined in its entire extent, from the pes hippocampi anteriorly to the fasciolar gyrus. The posterior limit of the hippocampus corresponded anatomically to the region of the pulvinar. The margins of the body of the hippocampus were defined superiorly as the choroidal fissure curving laterally along the medial boundary of the temporal horn, then medially along the gray matter of the hippocampus, including the medial-most part of the subiculum where the hippocampus joins the parahippocampal gyrus. Anteriorly the pes was distinguished from the amygdala, which is close to the antero-superior aspect of the pes near the tip of temporal horn and the uncus. Pixel points were summed from successive pictures to calculate structure volumes. HF volume images had 15 slices.

We also calculated left/right HF volume ratios, which are less sensitive to variations in head size, volumes of intracranial structures, and gender. A ratio less than 1 suggests left HF atrophy, and greater than 1 suggests right HF atrophy. We scanned 19 normal volunteers (10 women, 9 men). Because there was no difference between left and right HF volumes we calculated a mean normal bilateral HF volume measure.

Statistical analysis was performed on a personal computer using Systat 7.0 (SPSS Inc., Chicago, IL). Student’s t-tests, analysis of variance, and multiple regression analyses were performed for FS+, age at study, age at afebrile seizure onset, epilepsy duration, and the ratio between duration and age at study. Because men and women were equally represented among both patients and controls, we did not make separate gender comparisons.

Results.

Mean HF volume ipsilateral to the focus was lower than HF volume contralateral to the epileptogenic zone identified by ictal video-EEG monitoring: 2.73 ± 0.67 versus 3.24 ± 0.53 cm³ (paired t-test = 5.384; p < 0.0001). FS+ patients had smaller HF volume than did FS− patients ipsilateral but not contralateral to the epileptogenic zones, as well as a lower ipsilateral/contralateral ratio (table).

Compared with our normal controls (mean 3.29 ± 0.34 cm³), FS+ patients had small ipsilateral HF volume (p < 0.001) and a nonsignificant trend toward small contralateral HF volume (0.10 < p < 0.15). FS− patients had smaller HF volume only ipsilateral to the focus (p < 0.002). The FS+ group was too small to divide into men and women, but separate sex comparisons for the FS− group did not affect the results.

In simple regressions, epilepsy duration, but not age at scan or seizure onset, was significantly related to HF volume ipsilateral to the epileptogenic zone (R² = 0.172, F = 6.66, p < 0.02) and to the ipsilateral/contralateral ratio (R² = 0.154, F = 6.184, p < 0.02) (figures 1 through 3⇓⇓). HF volume contralateral to the focus tended to be smaller among FS+ than it was among FS− patients or controls (see table). When the effect of FS and duration was studied together, duration remained a significant variable for ipsilateral HF volume, although it had less weight (F = 4.997, p < 0.04) than did FS+ history (F = 9.691, p < 0.005). Neither age at afebrile seizure onset or age at study added significantly to the result. In a multiple regression, duration but not age had a significant effect on ipsilateral HF volume (figure 4). To help correct duration for the effect of age, we examined the relation of the ratio duration/age to HF volume. This variable had a significant effect on ipsilateral volume (R² = 0.144, p < 0.03) and the ipsilateral/contralateral ratio (R² = 0.183, p < 0.01), which was still present when FS were added to the regression. Among the patients with a history of complex FS, neither reported length nor latency to onset of afebrile seizures was correlated with HF volume.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. There is a significant decrease in hipocampal formation volume ipsilateral to the EEG focus with increasing duration of the seizure disorder (y = 3.203 − 0.023x; F = 6.66; p < 0.02).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Increasing duration of the seizure disorder is still associated with reduced hippocampal formation volume ipsilateral to the EEG focus even if patients with a history of complex or prolonged febrile seizures are excluded (y = 3.328 − 0.022x; F = 6.018; p < 0.03).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3. Increasing patient age did not have a significant effect on hippocampal formation volume ipsilateral to the epileptic focus (y = 3.191 − 0.013x; F = 1.335; p > 0.25).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4. Multiple regression plot of the interaction of the effect of age (x-axis) and duration (y-axis) on hippocampal formation (HF) volume ipsilateral to the epileptic focus (z-axis). Individual data points are shown. The smoother plane tilts down with increasing epilepsy duration and decreasing HF volume, but not with decreasing HF volume and increasing age (z = 2.81 + 0.023x − 0.04y; F = 4.242; p < 0.03).

Discussion.

Our study suggests that although patients with prolonged or complex FS are more likely to have HF volume loss, epilepsy duration has an additional effect. There is still a significant difference between the HF volume ipsilateral and contralateral to the epileptic focus in patients with no history of FS. Interestingly, the HF ipsilateral to the focus in FS− patients had the same volume as the contralateral HF in FS+ patients. There was a trend toward bilateral HF atrophy in the FS+ but not FS− patients, suggesting the effect of an early global insult. However, increasing epilepsy duration was inversely associated with ipsilateral but not contralateral HF volume, suggesting that any deleterious effect of persistent seizures is confined to the epileptogenic zone itself.

Other factors such as antiepileptic drug (AED) toxicity may play a role in the development of progressive HF atrophy. However, this effect would more likely be bilateral than more pronounced in the epileptic focus.

Several lines of evidence suggest that progressive damage may occur in epileptogenic zones in patients with TLE. The latent period frequently observed before seizure onset after an initial injury suggests an ongoing pathophysiologic process.13 Data from animal models show that in the dentate gyrus, a number of insults, including prolonged FS, head trauma, or encephalitis, may deafferent inhibitory neurons, disinhibiting the granule cell layer, leading to synchronous multilamellar discharges in response to cortical stimuli. Repetitive seizures may then lead to more extensive hippocampal damage.14 In rat dentate gyrus kindling models, intermittent, brief seizures induce both apoptotic death and proliferation of dentate gyrus neurons.15 In a kainic acid model of hippocampal sclerosis in the rat, mossy fiber sprouting increases progressively with longer survival after the lesion is created, suggesting an ongoing process.16 Moreover, an animal analogue of HF sclerosis can be induced by repeated kindled seizures in the rat. Cavazos et al.17 found that dentate gyrus and CA1 neuronal loss increased progressively with increasing seizure number, while other HF regions gradually became affected as well. No changes occurred in somatosensory cortex.

Human pathologic studies have also suggested increased damage with prolonged epilepsy duration.1,2 Babb et al.18 did not find an association with duration, but believed that this was due to the young age of most of their patients. Neuronal loss in CA1 and the prosubiculum in temporal lobe specimens resected at surgery was greater in patients who had had seizures for more than 20 years than it was in those with a shorter history.19 Patients with an “early risk factor” were reported to have lower hilar and granule cell densities in resected HF than were those without early risk.20 However, the patients also had longer epilepsy duration. HF cell loss may accompany extra-mesial temporal lesions, including gliomas, hamartomas, and heterotopias.21,22 Levesque et al.23 reported that heterotopias were associated with the most severe cell loss and that there was no relation to a history of febrile convulsions. These findings suggest that persistent seizures can lead to HF damage.

Progressive functional impairment in temporal lobe epileptogenic zones is supported by PET studies. We found a global reduction in glucose metabolism (CMRglc) related to epilepsy duration, but could not exclude the effects of age or other factors such as AEDs.24,25 Subsequent PET studies have shown an uncoupling between blood flow (CBF) and CMRglc in temporal lobe epileptic foci, with greater relative impairment of CMRglc.26,27 Moreover, the difference between the two measures increases with epilepsy duration, even though both appear to decline.28 The widening of the CBF–CMRglc mismatch was not related to a history of FS, onset age, or age at scan. AEDs, which have global effects on metabolism, are unlikely to be responsible for the process. Preliminary PET CMRglc data from a prospective study of children with new-onset seizures showed longer epilepsy duration in patients with hypometabolism.29

Previous vMRI studies of the relation of HF pathology to epilepsy duration have been contradictory. Spencer et al.4 reported that patients with mesial temporal seizure onset and HF atrophy on vMRI had significantly longer epilepsy duration than did those who had mesial temporal electrographic seizure onset but no atrophy, although age may not have been controlled for. Cendes et al.10 found that a history of FS, but not epilepsy duration, affected HF volume. Age at seizure onset was greater and epilepsy duration shorter than they were in our study. A preliminary study of newly diagnosed patients who had repeat MRIs after 1 year suggested that hippocampal sclerosis was found on a second study in several patients with normal initial scans.30 In a recent MRI study of children aged 8 to 33 months, VanLandingham et al.31 showed that prolonged or focal complex FS could be associated with acute HF injury. Progressive HF atrophy has been reported after status epilepticus both with and without persistent seizures, even when acute increases in T2 signal resolved.32,33 Fernandez et al.34 reported two families in which hippocampal malformations may have facilitated febrile convulsions and contributed to later epilepsy. Interestingly, several subjects in the study reported by VanLandingham et al.31 had evidence for HF pathology antedating the FS.

However, a minority of patients with uncontrolled CPS and mesial temporal sclerosis (MTS) have a history of complex or prolonged FS, or initial nonfebrile status. Kälviäinen et al.35 found an association between estimated seizure number and HF volume loss; in their study, both epilepsy duration and seizure number were correlated with increased T2 relaxation time.

The effects of TLE on the brain may extend beyond the primary epileptogenic zone. HF volume in our patients was reduced contralateral as well as ipsilateral to the epileptic focus, perhaps because of the severity of their epilepsy. Cook et al.36 reported that HF atrophy was widespread in patients with secondary generalized seizures. Ashtari et al.37 reported that left HF volume was significantly reduced in patients with right-sided EEG foci, but the reduction in right HF volume in patients with left-sided foci was not significant. Marsh et al.38 reported bilateral reductions of temporal neocortex and frontoparietal volume, as well as increased ventricular size, in patients with unilateral MTS. In a previous study, we found thalamic atrophy ipsilateral to the focus.39 Bilateral damage is found frequently in pathologic studies; these patients tend to have long epilepsy duration and severe seizures.1,2 HF damage, initially triggered by FS or another early insult, could be progressive in patients with uncontrolled TLE.