Abstract
INTRODUCTION
Alzheimer’s disease (AD) patients have an increased risk of epilepsy since seizure incidence is 6–10 times higher in patients than in age-matched healthy controls [1]. In mouse models with a heavy amyloid brain burden, AD-related epilepsy has been shown to be an early feature of the disease [2]. In these same models, epilepsy is amyloid-driven: amyloid-β (Aβ) peptides promote epileptogenesis. This leads to compensatory (antiepileptic) but also detrimental hippocampal remodeling in which synaptic plasticity, learning, and memory are impaired [2]. In sporadic AD, in vivo amyloid imaging has shown that the brain amyloid burden increases from 15 to 25 years before cognitive decline [3]. Amyloidopathy could thus be involved in early epileptogenesis in sporadic AD. Complementarily, animal models have shown that AD-related epileptogenesis is tau-dependent [2]. Tau pathology is initially located in the mesial temporal lobe structures when patients experience isolated anterograde amnesia defining a mild cognitive impairment of the amnestic type (aMCI), which is the most frequent prodromal syndrome of AD [4].
Accordingly, temporal lobe epilepsy could be an early feature of sporadic AD, occurring concomitantly with tau pathology in the hippocampal structures or even before cognitive decline [1, 5]. Only one study has specifically addressed the question of earlier seizures in sporadic AD by examining cases of epileptic aMCI: the authors showed that a pharmacosensitive partial non-convulsive epilepsy started at the time of cognitive decline and commonly relied on a temporal lobe focus [1]. In the present study, we describe a series of 13 cases of epilepsy occurring at a very prodromal stage of CSF-assessed AD.
MATERIALS AND METHODS
Patient selection
We searched the database of the Memory Center of the University Hospital of Strasbourg (n = 2,397) for all patients who presented cognitive decline and met the diagnosis criteria for MCI (n = 430 [258 aMCI+172 naMCI]). Among these patients we searched for subjects with a clinical diagnosis of epilepsy. 13 epileptic patients were thereof retrospectively selected. Medical history was retrospectively taken from patients and caregivers: we recorded the time of first cognitive changes and the time of the first spells with unresponsiveness and/or epileptic symptoms (including generalizations, see in Table 1). Epilepsy was diagnosed clinically, based on three main arguments: first, recurrent stereotyped and well-demarcated transients with ictal semiology suggestive of seizures according to current guidelines [6]; second, no arguments on the medical workup for an etiology other than seizure to account for the iterative transients; third, a more than 50% improvement in the frequency of spells with antiepileptic drugs. Prodromal AD was considered as causative of epilepsy because there were no other explanations on medical workup. Nevertheless, most of patients (84.6%) had cerebrovascular disorders (white matter changes and/or subcortical lacunes) that are considered as risk factors for late onset epilepsy [7]. Additionally, 3 patients (23.1%) had a history of mild traumatic brain injury and one of focal parietal subarachnoid hemorrhage (without secondary hemosiderosis).
Clinical measures
The whole non-epileptic MCI cohort (n = 417) included 245 aMCI. The demographics of the epileptic subjects were first compared to those of the 417 non-epileptic MCI and, second, to those of the 216 aMCI patients without verified CSF biomarkers. Third, the remaining 29 aMCI patients with positive CSF biomarkers were compared to the epileptic patients separately. All epileptic patients underwent clinical examination and Mini-Mental State Examination (MMSE) before and after antiepileptic drug (AED) treatment. MCI was diagnosed according to revised criteria: as daily living activities were preserved, patients showed shallow impairments (clinical dementia rating [CDR] = 0 to 0.5) on comprehensive cognitive testing (including the evaluation of memory, praxic abilities, visuo-construction, executive functioning, and language) [8]. During follow-up, neurodegenerative aphasic or neuro-visual symptoms occurred for patient 5 and 12 respectively (Table 1). For both these patients, the MMSE score was considered as non-representative of their real underlying global cognitive level and was therefore excluded from analysis.
Paraclinical measures
Standard blood tests, brain MRI (T1-T2-FLAIR-T2* weighted sequences) were systematically performed. Hippocampal atrophy was staged on coronal reconstructions with the visual scale of Scheltens ranging from 0 (no atrophy) to 4 (maximal atrophy) [9]. For all epileptic patients, an underlying Alzheimer neuropathology, thus defining MCI of the AD-type [10], was retained on CSF biomarkers showing results suggestive of amyloidosis (namely: decreased Aβ42 or Aβ42/Aβ40 ratio < 0.05 in the case of high amyloid producers) plus increased P-tau [10, 11]. The date of pathological CSF results was considered as the date of definite AD diagnosis.
Electroencephalogram (EEG)
Routine electroencephalography was performed in each patient, with the international 10–20 system for scalp electrodes on a 20-min session. Epileptic foci were defined as the regions of maximum electronegativity. Background slowing and/or epileptiform discharges (sharp waves or spikes) were considered as supportive arguments for the diagnosis of epilepsy [1].
Antiepileptic drug treatment
All epileptic patients were proposed an antiepileptic drug treatment when the diagnosis of epilepsy was clinically suspected. The choice of the drug was left to B.C. due to his experience in the management of elderly epileptic patients. The clinical response was measured on patient’s and caregiver’s diary at 3 and 6 months. Seizure reduction was achieved with relatively low doses of AED (Table 1). The 2 most used AED were lamotrigine and levetiracetam (Tables 1 and 2); all AEDs were well tolerated thus blood levels were not dosed. Seizure reduction > 50% was obtained in the following 3 months after initiation for 9 patients. Dose escalation was necessary for 4 patients (N° 1, 4, 8 and 11) and seizure alleviation was reached in the 6 months following seizure diagnosis. None experienced status epilepticus.
RESULTS
The demographics of epileptic and non-epileptic patients
The 13 epileptic prodromal AD patients were 3.1% of the whole MCI cohort (n = 430) and 5% of the aMCI group (n = 258). The 417 non-epileptic MCI subjects had the following stats: women = 52.6%, age (years, mean±SD) = 67,1 (±11.0), education (years) = 10.9 (±2.5), MMSE = 26.95 (±2.25). The demographics of the 216 aMCI patients without verified CSF biomarkers were: women = 48%, age = 67.96 (±10.5) years, education = 10.9 (±2.5), MMSE score = 27 (±2.2). The stats of the 29 patients with verified AD biomarkers were: women = 41%, age = 69.7 (±8.7), education = 11 (±3), MMSE score = 25.4 (±2.7). The epileptic patients tend to be older with a male predominance: women = 46.2%, age = 69.9 (±10.4) but they did not significantly differ for cognitive scores and education: MMSE = 25.9 (±2.3), education = 10.4 (±1.7).
Seizures onset in relation to cognitive decline, epilepsy diagnosis and prodromal AD diagnosis
Seizures retrospectively started at a mean (±SD) age of 63 (±12.8) years and were the first and isolated feature of the disease for a mean 2.7 years before cognitive decline at 65.7 (±12.2). AEDs were initiated when epilepsy was clinically suspected at 69.4 (±10.2; median = 68). Seizure reduction was obtained in the next 3 to 6 following months, resulting in the same age of an established diagnosis of epilepsy. Prodromal AD was diagnosed at the MCI stage almost 7 years after seizure onset (mean age = 69.9 [±10.4]) with a median MMSE score of 26 (mean = 25.9 [±2.3]) (Table 1).
Electroclinical characteristics of epileptic prodromal AD patients (Table 1)
Among the 13 patients, 84.6% had monthly (53.8%) or weekly (30.8%) simple partial seizures at the time of epilepsy diagnosis. Unresponsiveness and/or staring were present in 61.5% of patients. On EEG recordings, temporal slowing was unilateral in 38.4% of patients and bilateral in 23.1%; epileptiform discharges were additionally present over the left (15.4%) or right (7.7%) temporal lobe; 30.8% of patients had normal recordings. In combination with hippocampal atrophy (see next section below), these electro-clinical data are indicative of temporal lobe epilepsy (Tables 1 and 2) [1, 6]. The 13 treated patients achieved > 75% seizure reduction (seizure freedom for 11, including the 9 receiving lamotrigine monotherapy).
Neuropsychological characteristics of epileptic prodromal AD patients
At the time of diagnosis of prodromal AD, all epileptic patients had aMCI: they all had episodic memory impairments on cognitive tests but 69.2% had additional impaired domains (Tables 1 and 2). Brain imaging revealed mild bilateral hippocampal atrophy (median = 1 on scale of 0–4) [9] in all but 1 patient, white matter changes in 69.2% of patients and cerebro-vascular lesions in 38.5% (subcortical lacunes and microbleeds but no cortical lesions; Table 2). On CSF examination, increased P-tau (88.4±19.2 ng/ml; range: 65–117; N = 0–60) was consistently combined with an amyloidogenic process, as confirmed by a low CSF Aβ42 rate in 53.8% (mean = 314.6±76.5 ng/ml; range: 193–448; N = 500–1500 ng/ml) or by a low Aβ42/Aβ40 ratio (mean = 0.04±0.0038; range: 0.036–0.045; N > 0.05) in the remaining 46.2% with normal CSF Aβ42 (range: 635–1494 ng/ml). After AED treatment the MMSE score remained stable at a mean 26.3±2.4 (after a median 10.5 months of follow-up) with subjective cognitive relief for only 2 patients (15.4%).
DISCUSSION
We have replicated most of the electro-clinical data of Vossel et al. [1] (Table 2) but our patients had a longer delay between seizure onset and neurodegenerative diagnosis (6.9 years versus 3.9 years in [1]) and had more frequent complex partial seizures (Table 2). A retrospective bias in the past medical history interview and a difference between the diagnostic criteria used (CSF versus clinically defined AD) could account for such discrepancies. The retrospective bias also explains the delay between the onset of seizures and the established diagnosis of epilepsy (6.4 years, Table 1). Furthermore, simple partial seizures are difficult to identify in the elderly, because several other causes can be responsible for short transients of non-specific semiology [12]. Thus the diagnosis of epilepsy is long to be established in older subjects. In fact, 7 of our patients (i.e., patients 1, 3, 4, 7, 8, 9, and 11) first had subjective epileptic transients that were initially overlooked. The epileptic picture became clearer when complex partial seizures and/or generalizations emerged (Table 1). Our patients had less frequent epileptiform EEGs (23.1%) than those of Vossel et al. (60% = 6 of the 10 recorded subjects): it is probably because we only used standard electroencephalography whereas most of Vossel’s subjects had extended long-term monitoring or serial-EEGs when routine EEGs were negative.
The patients were treated for epilepsy before being diagnosed with incipient AD but the dates of both these diagnoses are close (69.4 and 69.9 years, respectively). One explanation is that the newly diagnosed epilepsy was combined, in these elderly patients, with neurodegenerative features (i.e., hippocampal atrophy and memory impairment) suggestive of possible underlying AD [10]. Another reason is that patients still complained of their cognition in spite of good seizure response to AEDs: this led us to CSF amyloid biomarkers examination [5]. Hence, our diagnosis of prodromal sporadic AD is reliable because it is based on positive CSF analysis [10].
Our results add new data to characterize the poorly described epileptic prodromal AD [1, 5]. First, we provide more detailed cognitive profiles (impaired domains other than episodic memory in ≈ 70%, Tables 1 and 2). Second, we describe mild hippocampal atrophy combined with white matter changes and vascular lesions that are probably related to the vascular risk factors of our patients. Third, our data comprise the outcome under AED treatment: interestingly, the MMSE score remained stable in spite of the poor subjective improvement (Tables 1 and 2). Compared to the non-epileptic MCI group, our patients appeared older with a mean age of 69.9 versus 67.1. They had slightly lower MMSE scores (25.9 versus 26.95, respectively). These data must be cautiously considered due to our retrospective methodology and the small size of our cohort. Nonetheless, limited seizures (i.e. simple or complex focal seizures) have a limited effect on global cognitive abilities and on memory [13, 14].
Since Vossel et al.’s patients [1] and ours share many similarities (Table 2), they together delineate a “new” epileptic syndrome: it consists in drug-responsive non-convulsive temporal or temporo-frontal lobe epilepsy that starts around the age of 60 and remains isolated for 1 to 10 years. Thereafter, memory complaints occur and lead to the diagnosis of sporadic AD (3 to 5 years later) at the MCI stage (MMSE score ranging from 23 to 30 with mild hippocampal atrophy on imaging). Treatment with anticonvulsants does not generally provide any subjective cognitive relief [1], despite seizure reduction and MMSE score stability over the first year of follow-up. In our work, seizures were isolated for a mean period of more than 2.5 years: this suggests a peculiar and as yet unrecognized variant of sporadic AD that could be called the “epileptic variant of Alzheimer disease” (evAd). From our point of view, the evAd concerns patients with overt clinical seizures that precede any other symptoms of sporadic AD. We cannot be sure that it also describes prodromal AD patients with subclinical epileptiform activity because, by definition, they do not complain of recurrent disturbing spells [1].
Due to its methodology our work does not allow any conclusions on the links between epilepsy and AD severity [1] and on the potential benefits of AEDs on cognition (i.e., improvement and/or prevention of a subsequent worsening). We are also unable to confirm whether the high Aβ rates and/or the vascular lesions are significantly involved in the epileptogenesis of our patients [2]. Prospective, longitudinal and case-controlled studies are now needed to test these hypotheses and to confirm the interest of describing the evAd. In one hand, such studies would answer if a presenile or senile epilepsy (especially of temporal type), starting before a cognitive impairment, is the cause of this alteration. In the other hand, they may indicate that epilepsy is an aggravating factor in AD for which AED treatment has a positive impact on cognition.
Disclosure Statement
Authors’ disclosures available online (http://j-alz.com/manuscript-disclosures/15-0096r2).