The Connection Nobody Told You About: Epilepsy and Dementia

Most of my blog topics come from patients. A question asked across the exam room in my resident clinic, a family navigating something I had not written about yet, a clinical moment that stays with me on the drive home. This week was different.

One of my favorite attendings — an epileptologist whose clinical instincts I deeply respect — mentioned a term in passing during a conversation: epilepsy-related dementia. I nodded, but when I got home I realized I had never given this relationship the serious attention it deserves. We see seizure patients constantly in neurology, both inpatient and in clinic — young adults, middle-aged patients, older patients with new-onset episodes — and almost never do I find myself connecting those encounters explicitly to the dementia prevention framework I spend so much time thinking about and passionately researching.

What I found in the literature is not a peripheral association. Which I suppose I have to confess really surprised me.  It is a robust, bidirectional relationship between epilepsy and dementia that has been quietly building in the research base for years — and that has real, actionable implications for patients living with epilepsy right now.

The Numbers Are Not Small

Let me start with the epidemiology, because the scale of this relationship is important enough to be stated explicitly.

A 2024 systematic review and meta-analysis published in Epilepsy & Behavior pooled data across multiple large studies and found that epilepsy is associated with a nearly threefold increased risk of developing dementia compared to individuals without epilepsy — a risk ratio of 2.83. Among those with epilepsy, the pooled prevalence of eventual dementia was 17%, and of Alzheimer's specifically, 15%.

The other direction is equally striking. People with Alzheimer's disease have a 2.3-fold greater risk of developing seizures or epilepsy compared to the general population. Among early-onset Alzheimer's patients specifically, the five-year risk of developing epilepsy is more than four times that of age-matched controls. And subclinical epileptiform activity — abnormal electrical discharges detectable only on EEG, without overt clinical seizures — has been found in 20 to 50% of Alzheimer's patients across multiple studies.

Two common neurological conditions, each dramatically increasing the risk of the other. And yet this conversation is almost never part of routine epilepsy care (in my experience). 

What Is Actually Happening in the Brain

To understand why this relationship exists, you have to look at what epilepsy and Alzheimer's disease do to the same brain structures — and how disturbingly similar the two processes look at the molecular level.

Both conditions attack the hippocampus and medial temporal lobe with particular preference. Both produce amyloid accumulation and tau pathology (I truly had not appreciated this point) — the hallmark proteins of Alzheimer's disease. Both cause hippocampal sclerosis, the scarring and cell loss in hippocampal tissue that is both a consequence and a driver of further neurological dysfunction. Much more commonly, Alzheimer’s is connected to hippocampal atrophy (shrinkage), but I have reviewed research now that more ties it to scarring specifically as well.  This is not coincidence. It reflects shared pathophysiology that runs deeper than the clinical diagnoses suggest.

The specific mechanism connecting seizure activity to accelerated neurodegeneration is now reasonably well characterized. When neurons fire abnormally and repeatedly during a seizure, they dramatically increase their metabolic demands and release large quantities of glutamate — the brain's primary excitatory neurotransmitter. This glutamate surge is directly toxic to surrounding neurons and appears to accelerate the production and aggregation of both amyloid-beta and hyperphosphorylated tau. Chronic seizures, in this framework, are not just electrical events. They are neurotoxic insults that leave a molecular residue.

The reverse mechanism is equally well established. The progressive accumulation of amyloid plaques and tau tangles in Alzheimer's disease disrupts the excitatory-inhibitory balance of neural networks by impairing the function of inhibitory interneurons — the brain's internal braking system. As those brakes fail, cortical hyperexcitability develops, lowering the seizure threshold and making spontaneous epileptiform activity progressively more likely. The two diseases, in other words, actively create the conditions for each other.

Temporal lobe epilepsy deserves specific mention here because its pathological overlap with Alzheimer's is particularly significant. The hippocampus and entorhinal cortex — the epicenter of temporal lobe epilepsy — are also the regions where Alzheimer's pathology begins and is most severe. Patients with mesial temporal lobe epilepsy show amyloid deposition, tau accumulation, and hippocampal sclerosis patterns that mirror early Alzheimer's changes. Some researchers have begun to ask whether temporal lobe epilepsy and early Alzheimer's may in some patients represent different clinical expressions of converging pathological processes rather than truly separate diseases.

There is also a genetic thread worth noting. The APOE-E4 allele — the strongest known genetic risk factor for Alzheimer's — appears to confer elevated risk for both conditions. A 2025 study in Scientific Reports using multi-task learning to identify shared genetic risk identified the APOE-TOMM40-APOC1 cluster and several other loci as common risk architecture for both late-onset epilepsy and Alzheimer's disease. The shared genetic substrate further supports the idea that these are not independent processes running in parallel but overlapping pathological cascades with common upstream drivers.

The Late-Onset Epilepsy Signal

One piece of this puzzle deserves particular emphasis for patients and families who may not have connected these dots.

Late-onset unexplained epilepsy — new-onset seizures after age 55 for which no structural, vascular, or metabolic cause can be identified — accounts for up to 25% of all new epilepsy diagnoses and affects approximately 50,000 new patients in the United States each year. Research now consistently shows that this group carries a markedly elevated risk of subsequent dementia: studies suggest that 10 to 25% of individuals with late-onset unexplained epilepsy will develop dementia within three to four years of their first seizure.

A 2026 personal view published in The Lancet Neurology went further, arguing that late-onset epilepsy may not merely be a risk factor for Alzheimer's but may in some cases represent a prodromal symptom within the Alzheimer's continuum — a neurological signal that the underlying disease process is already underway, preceding overt cognitive symptoms by years.

This reframes the clinical significance of a new seizure in a 60 or 65-year-old patient without an obvious cause. It is not simply an epilepsy diagnosis to be managed with medication. It is a potential early warning that the neurodegenerative process deserves evaluation — including consideration of amyloid and tau biomarker testing, and a proactive conversation about cognitive monitoring going forward.  That is how I plan to manage it from this point forward.

What Patients With Epilepsy Can Actually Do

The research on modifiable risk within this relationship is still developing, but several principles are supported well enough to act on now.

Seizure control is neuroprotective. This may be the single most important clinical message. A pilot study examining cognitive outcomes in adult epilepsy patients found that a seizure frequency of once per month or more was strongly associated with a dramatically increased risk of mild cognitive impairment — with an odds ratio exceeding 100 in multivariate analysis. Epilepsy duration and the number of medications used were not independently significant; it was the seizures themselves that drove the cognitive risk. Getting to seizure freedom, or the lowest achievable seizure frequency, is not just a quality-of-life goal. It is a brain health goal.

Medication choice matters. A meta-analysis published in Frontiers in Neurology found that older first-generation anti-seizure medications — phenobarbital, phenytoin, carbamazepine, valproate — were associated with elevated dementia risk compared to newer agents. Levetiracetam (Keppra) has shown preclinical evidence of cognitive protection and has been investigated specifically in Alzheimer's patients for its potential to stabilize hyperexcitable neural networks. This does not mean that older medications should be reflexively discontinued — many patients are well-controlled on them — but it does mean that the cognitive profile of anti-seizure medication choices is a legitimate part of the treatment conversation, particularly for middle-aged and older patients.

Sleep is a shared vulnerability. Epileptiform discharges are more frequent and more damaging during sleep disruption, and impaired sleep independently accelerates both seizure frequency and the amyloid accumulation that drives Alzheimer's. For epilepsy patients, sleep hygiene and screening for obstructive sleep apnea are not peripheral recommendations. They sit at the intersection of seizure control and dementia prevention simultaneously.

Vascular risk factor management. Hypertension, diabetes, and metabolic syndrome are shared risk factors for both epilepsy and dementia. Managing them aggressively improves outcomes in both directions — reducing seizure susceptibility and reducing the cerebrovascular pathology that accelerates neurodegeneration.

Cognitive monitoring should be routine. Patients with epilepsy — particularly those with temporal lobe involvement, older age of onset, or frequent breakthrough seizures — should have periodic cognitive screening as part of their neurological follow-up. This is not yet standard practice in most neurology clinics where patients are being managed for their epilepsy. It should be.

Takeaways

Epilepsy and dementia have a bidirectional relationship supported by robust epidemiological and molecular evidence. Epilepsy nearly triples the risk of dementia; Alzheimer's disease more than doubles the risk of seizures.

The mechanism is molecular, not coincidental. Chronic seizure activity accelerates amyloid and tau accumulation through glutamate toxicity and metabolic stress. Alzheimer's pathology lowers seizure thresholds by impairing inhibitory interneuron function. Each disease actively promotes the other.

Late-onset unexplained epilepsy after age 55 is a potential early Alzheimer's signal. Up to 25% of affected individuals develop dementia within four years. This diagnosis should prompt biomarker evaluation and proactive cognitive monitoring, not just anti-seizure management.

Seizure control is the most actionable neuroprotective target. Frequent breakthrough seizures are strongly associated with cognitive decline. Achieving the lowest possible seizure burden — through optimized medication, lifestyle, and sleep management — is a dementia prevention strategy.

Medication choice, sleep, and vascular risk management are all modifiable. The cognitive profile of anti-seizure medications deserves attention, particularly in older patients. Sleep disruption amplifies both seizure risk and neurodegeneration. Vascular health is a shared upstream driver of both conditions.

Citations and References

Stewart D, Johnson EL. The Bidirectional Relationship Between Epilepsy and Alzheimer's Disease. Current Neurology and Neuroscience Reports. 2025. https://doi.org/10.1007/s11910-025-01404-y

Vossel K, et al. Epileptic activity in Alzheimer's disease: emerging insights and therapeutic implications. The Lancet Neurology. 2026. https://doi.org/10.1016/S1474-4422(25)00425-9

Tang Y, et al. The crosstalk between epilepsy and dementia: a systematic review and meta-analysis. Epilepsy & Behavior. 2024. https://doi.org/10.1016/j.yebeh.2024.109640

Lam AD, et al. Late-onset unexplained epilepsy as a risk factor for cognitive impairment and dementia (ELUCID study protocol). medRxiv. 2025. https://doi.org/10.1101/2025.06.16.25329698

Liang CS, et al. Anti-seizure medication exposure and the risk of dementia: A meta-analysis of observational studies. Frontiers in Neurology. 2023. https://doi.org/10.3389/fneur.2023.1133816

Multi-task learning identifies shared genetic risk for late-onset epilepsy and Alzheimer's disease. Scientific Reports. 2025. https://doi.org/10.1038/s41598-025-32329-8er