Abstract
Joutsa J, Rinne JO, Niemi KJ, Karrasch M, Parkkola RK, Saunavaara J, Helin SP, Hermann BP, Sillanpää M. Neurology. 2025;104(3):e210303. doi: 10.1212/WNL.0000000000210303. Background and objectives: Previous research has demonstrated increased brain amyloid plaque load in individuals with childhood-onset epilepsy in late middle age. However, the trajectory of this process is not yet known. The aim of this study was to determine whether individuals with a history of childhood-onset epilepsy show progressive brain aging in amyloid accumulation in late adulthood (Turku Adult Childhood-Onset Epilepsy study, TACOE). Methods: Adults from a prospective population-based cohort of individuals with childhood-onset epilepsy, originally recruited 1961-1964, together with matched controls, were scanned with [11C]PIB PET twice: after at least 50 years (TACOE-50) and again after at least 55 years (TACOE-55) from the diagnosis. Results: At TACOE-55, 31.4% (11/36, mean age 63.3 years, 52.8% female) of individuals from the epilepsy group and 11.4% (4/35, 63.1 year, 54.3%) of controls had a visually abnormal [11C]PIB scan (p = 0.039). At TACOE-55, cortical brain [11C]PIB uptakes were higher and increased more from TACOE-50 in the epilepsy compared with the control group (p < 0.05). In voxelwise whole-brain analyses, the epilepsy group showed significantly higher and more widespread brain amyloid accumulation (pFWE < 0.05). Discussion: The results demonstrate that childhood-onset epilepsy is associated with an earlier age at onset of amyloidosis and greater progressive amyloid accumulation in late adulthood.
Fonseca E, Lallana S, Ortega G, Cano A, Sarria-Estrada S, Pareto D, Quintana M, Lorenzo-Bosquet C, López-Maza S, Gifreu A, Campos-Fernández D, Abraira L, Santamarina E, Orellana A, Montrreal L, Puerta R, Aguilera N, Ramis M, de Rojas I, Ruiz A, Tárraga L, Rovira À, Marquié M, Boada M, Toledo M. Epilepsia. 2024 Dec;65(12):3664–3675. doi: 10.1111/epi.18142. Objective: Pathological amyloid-β (Aβ) accumulation and hyperphosphorylated tau proteins have been described in resected temporal lobe specimens of epilepsy patients. We aimed to determine cerebrospinal fluid (CSF) Aβ1–42 and p181-tau levels and cerebral Aβ deposits on positron emission tomography (Aβ PET) and correlate these findings with cognitive performance in adults with drug-resistant temporal lobe epilepsy (TLE). Methods: In this cross-sectional study, we enrolled individuals with drug-resistant TLE who were 25–55 years old. Each participant underwent 18F-flutemetamol PET, determination of CSF Aβ1–42, p181-tau, and total tau, and a comprehensive neuropsychological assessment. We evaluated normalized standard uptake value ratios (SUVRs) for different brain regions on Aβ PET. Results: Thirty patients (mean age = 41.9 ± SD 8.1 years, 57% men) were included. The median disease duration was 9.5 (interquartile range = 4–24) years. Twenty-six patients (87%) had a clinically significant cognitive impairment on neuropsychological evaluation, 18 (69%) of the amnesic type. On Aβ PET, high uptake was observed in both mesial temporal regions (ipsilateral: SUVR z-score = .90, 95% confidence interval [CI] = .60–1.20; contralateral: SUVR z-score = .92, 95% CI = .57–1.27; p < .001), which was higher when compared to SUVR z-scores in all the remaining regions (p < .001) and in the ipsilateral anterior cingulate (SUVR z-score = .27, 95% CI = .04–.49, p = .020). No significant deposition was observed in other regions. Seven patients (23%) had low Aβ1–42 levels, and two (7%) had elevated p181-tau levels in CSF. Higher p181-tau levels correlated with poorer verbal fluency (R = −.427, p = .044). Significance: Our findings reveal a considerable Aβ deposition in mesial temporal regions and ipsilateral anterior cingulate among adults with drug-resistant TLE. Additionally, abnormal CSF Aβ1–42 levels were observed in a significant proportion of patients, and p181-tau levels were associated with verbal fluency. These results suggest that markers of neuronal damage can be observed in adults with TLE, warranting further investigation.
Commentary
Persons with epilepsy are at risk of accelerated aging and cognitive decline. Longitudinal neuroimaging studies show accelerated patterns of atrophy, 1 and epidemiologic studies indicate a fourfold increased risk of developing dementia. 2 Studies of surgical specimens from epilepsy patients provided some clues with regard to the pathophysiology, showing elevated levels of amyloid beta and its precursor proteins. 3 Similarly, rat models of mesial temporal lobe epilepsy (TLE) exhibit activation of amyloidogenic pathways and tau hyperphosphorylation, similar to those seen in Alzheimer's disease (AD). 4 This evidence suggests a close link between epilepsy and the neurodegenerative proteins amyloid and tau. We now know that this relationship consists of a vicious cycle with seizures begetting amyloid and tau, which in turn contribute to hyperexcitability. 5
To date, studies exploring this relationship have been limited to either surgical specimens or cohorts with late-onset epilepsy. The implications of this relationship for individuals with early onset epilepsy were previously unknown, which is where the two new studies of interest come in.6,7
In the first study, Joutsa et al 6 describe the amyloid positron emission tomography (PET) findings of their unique Turku Adult Childhood-Onset Epilepsy (TACOE) study, TACOE cohort. This is a cohort of individuals in Finland with childhood-onset epilepsy who were prospectively followed from the 1960s and invited to participate in an amyloid [11C]PIB PET study and cognitive testing 50 years after study initiation; this was the TACOE-50 study. The study also included controls from the general population. The TACOE-50 study showed that individuals with childhood-onset epilepsy had higher amyloid burdens in the frontal lobes as compared to their peers. 8 Both epilepsy and control participants were invited again for a repeat amyloid PET and cognitive testing an average of 7.2 years later for the TACOE-55 study. The epilepsy cohort consisted of 36 participants with an average age of 63.3 years, 31.4% with an apolipoprotein ε4 (APOEε4) allele, 58.3% with a history of generalized seizures, and 22.2% with active epilepsy. The epilepsy type included 52.8% with idiopathic epilepsy and 47.2% with unknown etiology. The percentage of epilepsy patients with a positive scan was higher in the epilepsy group versus controls (30.6% vs. 11.4%) based on visual assessments. In the epilepsy cohort, there was a more widespread pattern of PIB uptake compared to the prior scan in TACOE-50. APOEε4 carriers had higher whole-brain PIB uptake. Amyloid positivity was associated with increasing numbers of impaired cognitive tests. Notably, the type of epilepsy and status (active vs. inactive) were not associated with the PIB findings.
In the second study, Fonseca et al recruited 30 individuals with drug-refractory TLE who underwent an amyloid PET with the 18F-flutemetamol (Vizamyl) tracer as well as cerebrospinal fluid (CSF) testing for Aβ1–40, Aβ1–42, p181-tau, total tau (t-tau), and cognitive testing. The average age of this cohort was 43.5 years, epilepsy duration 9.5 years, 17% with tonic–clonic seizures, and 20% APOEε4 carriers. The cohort did not have a control group, but the investigators used established cut-offs for the CSF biomarkers and a normative database of PET findings of >100 healthy individuals provided by the PET manufacturer. The PET images showed elevated tracer levels in the bilateral mesial temporal regions and the cingulate region ipsilateral to the epileptogenic focus. The parietal and occipital regions showed lower tracer uptake. On CSF testing, 23% had low abeta levels, and 7% had elevated p-tau181 levels but none had both. High levels of CSF ptau181 were associated with lower verbal fluency performance, no other correlations were found between the other CSF markers or PET findings and cognitive outcomes. APOEε4 carriers did not have higher PET tracer uptake.
Before discussing the relevance of the findings, one should note a few caveats: (1) both cohorts studied were relatively small in the 30–36 and heterogenous with regard to type of epilepsy; Idiopathic versus focal versus unknown in the case of TACOE, and with regard to pathology, with or without mesial temporal sclerosis in the case of the TLE study. (2) Due to technical limitations, the investigators in the TACOE study could not quantify the absolute increases in amyloid within subjects. (3) It is also technically challenging to determine amyloid binding in small regions such as the hippocampus especially in the setting of atrophy, and due to potential off-target binding effects because of nearby structures such as the choroid plexus.
Let us look at these studies through an Alzheimer's lens. In AD, the amyloid cascade hypothesis 9 posits that amyloid is the prerequisite protein that leads to tau hyper-phosphorylation and accumulation, followed by neuronal death, atrophy, and cognitive decline. In Alzheimer's, cortical areas are the first areas involved in amyloid buildup, and the hippocampus is only involved in later stages, so some of the PET findings seem epilepsy specific, given that in TLE, it was mostly limited to the hippocampi. The amyloid proteins are also believed to behave like prions and spread to other brain regions by trans-synaptic spread. This could explain the cingulate findings in the second study because the cingulate and the hippocampus are components of the limbic circuit and are highly interconnected. This could also explain the progressive accumulation of amyloid in TACOE-55; once it was established at an early age, it continued accumulating and spreading. We also know that the two most common factors associated with amyloid accumulation are age and having at least one copy of the APOEε4 allele. These findings were replicated somewhat in the TACOE study.
The studies raise many questions that we must consider before labeling the findings as a “bad news” headline for anyone with a history of epilepsy. Both studies did not identify epilepsy-related risk factors for the presence of amyloids, such as drug resistance, epilepsy duration, or the frequency and severity of seizures. It is difficult to believe that someone who had epilepsy, which was self-limited and medication-responsive as a child, is at risk for early onset amyloid deposition. Could there have also been a contribution from subclinical epileptiform activity as well? How many and what type of seizures are enough to trigger the amyloid cascade? What does this also mean for epilepsy surgery? Should we view the TLE study as another indicator that we must remove the epileptogenic lesion as soon as possible, not only because of kindling concerns but because of the possibility of increasing amyloid burden and spread? Should we consider treating a 30-year-old with cognitive impairment and a positive hippocampal amyloid PET with amyloid antibodies, and are there epilepsy-specific PET thresholds to consider a PET-positive? Can we identify patterns and thresholds that will identify whether someone is at risk of accelerated cognitive decline or AD?
With the rapid advances in biomarker testing, we can phenotype our epilepsy patients better. However, we will need more studies to understand how aging-related proteins affect the course of epilepsy and the aging process in individuals with epilepsy. Additionally, we must remember that amyloid positivity is not destiny; many amyloid-positive individuals do not go on to develop AD. 10 This highlights the need for more studies looking at aging and brain resilience in individuals with a history of epilepsy and how we, as physicians, can facilitate this resilience.
