Abstract
Progression of Gray Matter Atrophy in Seizure-Free Patients with Temporal Lobe Epilepsy
Alvim MKM, Coan AC, Campos BM, Yasuda CL, Oliveira MC, Morita ME, Cendes F. Epilepsia 2016;57:621–629. doi:10.1111/epi.13334.
OBJECTIVES: To investigate the presence and progression of gray matter (GM) reduction in seizure-free patients with temporal lobe epilepsy (TLE). METHODS: We enrolled 39 consecutive TLE patients, seizure-free for at least 2 years—20 with magnetic resonance imaging (MRI) signs of hippocampal sclerosis (TLE-HS), 19 with normal MRI (TLE-NL), and 74 healthy controls. For longitudinal analysis, we included individuals who had a second MRI with minimum interval of 18 months: 21 patients (10 TLE-HS, 11 TLE-NL) and 11 controls. Three-dimensional (3D) T1-weighted images acquired in 3 Tesla MRI were analyzed with voxel-based morphometry (VBM). The images of patients with right-sided interictal epileptogenic zone (EZ) were right–left flipped, as well as a comparable proportion of controls. Cross-sectional analysis: The patients’ images from each group were compared to controls to investigate differences in GM volumes. Longitudinal analysis: The first and second images were compared in each group to look for decreased GM volume. RESULTS: Cross-sectional analysis: Patients with TLE-HS had diffuse GM atrophy, including hippocampus and parahippocampal gyrus, insula, frontal, and occipital lobes ipsilateral to EZ, bilateral thalamus and contralateral orbitofrontal gyrus, and caudate. In contrast, TLE-NL group did not present significant differences compared to controls. Longitudinal analysis: TLE-HS presented progressive GM reduction in ipsilateral insula and occipital lobe, contralateral motor area, and bilateral temporal and frontal lobes. TLE-NL had GM progression in ipsilateral hypothalamus and parietal lobe, contralateral cerebellum, and bilateral temporal lobe. Controls did not show changes in GM volume between MRIs. SIGNIFICANCE: Diffuse extrahippocampal GM atrophy is present in seizure-free patients with TLE-HS. In addition, there is progressive GM atrophy in patients with and without HS. These results demonstrate that not only ongoing seizures are involved in the progression of GM atrophy. An underlying pathologic mechanism could be responsible for progressive brain volume loss in TLE patients even in seizure-free periods.
Commentary
The potential for epilepsy to be a progressive disorder has been debated for decades and has received considerable attention from the neuroimaging research community. In 2003, in Epilepsy Currents, Greg Cascino reviewed an article entitled “Progressive Neocortical Damage in Epilepsy,” which concluded that in focal epilepsy “ongoing cerebral atrophy may be widespread and remote from the putative epileptic focus” (1, 2). The investigators evaluated 122 patients with chronic epilepsy, 68 newly diagnosed patients, and 90 controls by using serial MRI scans obtained 3.5 years apart and found that 54% of those with chronic epilepsy, 39% with newly diagnosed epilepsy, and 24% of controls showed progressive volume loss. Of importance, risk factors for progressive neocortical atrophy were age and multiple antiepileptic drug exposure (2).
Alvim et al. recently reported a study of 74 healthy controls and 21 patients with temporal lobe epilepsy (TLE) (10 with hippocampal sclerosis ([TLE-HS] and 11 with normal MRI [TLE-NL]), who had been seizure-free for at least 2 years and who had received a second MRI at a minimum interval of 18 months. The repeated 3 Tesla MRI analyzed with voxel-based morphometry found that progressive extrahippocampal gray matter (GM) atrophy occurs in TLE-HS and TLE-NL patients, but at baseline imaging, GM atrophy was only present in the TLE-HS group. This study provides additional, important evidence for progressive cerebral changes in TLE. However, some considerations remain regarding the role of seizures in progressive injury, and also whether hippocampal networks are uniquely vulnerable to ongoing injury. The fact that the entire sample had experienced recurrent seizures for more than 10 years prior to the median 3- to 4-year period of remission before the index imaging study may indicate that many years of recurrent seizures can create an injurious process that does not end immediately upon seizure cessation. The significance of the minimal difference in baseline GM loss between TLE with and without HS remains uncertain because the two groups in this study significantly differed in age (51 vs 39 years) and number of antiepileptic drugs (2 vs 1), which were the two strongest predictors of progressive cerebral atrophy in the study by Liu et al. (2). The HS group in the Alvim et al. study was older and had a higher antiepileptic drug burden; it is likely that the TLE-NL group would continue to have GM loss during an additional 12 years of follow-up. Also, as noted by the authors, the small number of subjects in the TLE groups limited the statistical power to determine a significant difference in GM loss between those with HS and those with normal MRI.
The observation using MRI of progressive, widespread GM loss in patients with TLE has potentially problematic ramifications from a functional perspective, suggesting that cognitive and behavioral deterioration may follow the structural changes. However, the literature seems indeterminate regarding long-term cognitive and psychiatric outcomes in TLE, and potential interpretations of reported findings can be complex. For example, Helmstaedter and Elger performed a cross-sectional study of verbal learning and memory in 1,000 healthy controls (aged 6–80 years) and 1,156 patients with chronic TLE (aged 6–68 years) with a mean duration of seizures of 18 years (3). Although the learning peak (i.e., crossover to cognitive decline) was seen earlier in TLE patients (age 17 years) than for controls (age 24 years), the rate of decline in cognition ran in parallel between groups. In fact, the rate of decline after age 55 seemed to slow in TLE patients, whereas it increased in healthy controls. These results could suggest that long-term cognitive decline may not be as severe as the progressive GM atrophy could presuppose. Another neuroimaging study reported a paradoxical discordance between cerebral atrophy and psychiatric symptoms. Hecimovic et al. found that TLE patients with the most severe HS had significantly less depression than those with mild to moderate sclerosis, suggesting that a critical mass of abnormal hippocampal neurons is necessary to cause dysfunction in limbic networks (4). This hypothesis was further supported by the observation that decreased connectivity of temporal and inferior frontal regions in TLE is associated with less depression (5) and is consistent with prior anatomic studies that demonstrated that the densest connectivity of Broadmann's Area 25 (involved in the experience of depression symptoms) is with the anterior hippocampus (6).
Alvim et al. provide important confirmation of the progression of diffuse GM loss in TLE as well as support for the concept that factors other than seizures may play an important role in the process of cerebral injury. They provide another noteworthy example of the contributions of advanced neuroimaging techniques to the neuroscience of epilepsy. However, our understanding of the causes of cerebral loss and its association with cognitive decline and psychiatric problems remains, well…still somewhat gray.