Subscribe to eTOC

A Population-based Analysis Links Epilepsy to Dementia

Article In Brief

Epilepsy patients may have an increased likelihood of developing dementia compared with people who do not have the condition, according to a population-based analysis in Wales. Experts say these results suggest that clinicians should screen for signs of cognitive decline in their patients with epilepsy and monitor medications that may contribute to the increased risk for dementia.


Predicted hazard of dementia, Alzheimers disease, and vascular dementia in people with and without epilepsy, fee of dementia at age 60.

Individuals with epilepsy had a higher likelihood of developing dementia than people who did not have the disorder, according to a population-based analysis published in the July 17 online edition of Neurology.

The retrospective study reviewed medical records and ICD-10 codes from a large database comprising general practice consultations, prescriptions, hospital admissions, and death records for more than a half-million residents of Wales, UK.

The study authors reported that, compared with people who did not have epilepsy, those with the disorder were 2.5 times more likely to have dementia, 3.1 times more likely to have vascular dementia (VaD), and 1.6 times more likely to have Alzheimer's disease (AD). In addition, they reviewed exposure to certain antiepileptic drugs (AEDs) and determined that certain AEDs were associated with a higher risk for dementia.

Epileptologists who reviewed the paper for Neurology Today said the findings suggest that clinicians should screen for signs of cognitive decline in their patients with epilepsy. As well, they should stay vigilant about monitoring medications that may contribute to that increased risk for dementia.

Study Details, Findings

For their analysis, the research team examined the medical records of residents, aged 60 and older, to dementia diagnosis, death, or till the end of the study follow-up in June 2018. The median year at the start of follow-up (60th birthday) was 2006. Residents were excluded if they had a dementia-related health record before their 60th birthday or if they had no diagnosis of epilepsy.

Among findings, 16,700 people in the cohort had dementia, of which 6,916 were classified as having AD and 5,005 VaD. In addition, 1,060 people had codes for both AD and VaD, and 5,839 had unspecified dementia and could not be subtyped. Of the 19,150 (3.4 percent) people with epilepsy, 3,274 had an epilepsy diagnosis on or before their 25th birthday.

Within 21, 25, and 26 years of follow-up from their 60th birthday, the researchers predicted that 20 percent of people with epilepsy would have a diagnosis of any dementia, AD, and/or VaD. Within 25 and 31 years, the likelihood increased to 40 percent for any dementia and VaD, respectively. Within 31 years, the likelihood was 31 percent for AD.

Controlling for confounding factors such as smoking, hypertension, social deprivation, alcohol, diabetes, head injury, and depression, did not change the risk of dementia for people with epilepsy. However, the risk of VaD but not AD was attenuated with a diagnosis of stroke.

“Patients with a history of stroke might be more likely to be diagnosed as having VaD in routine clinical practice, even if some concurrent AD is present in reality,” they wrote.

The risk for dementia, AD, or VaD was similar in people diagnosed with epilepsy at age 25 years or younger as it was with people diagnosed later in life, according to the findings.

“It was an interesting and somewhat surprising finding that those diagnosed with epilepsy in early life had a similar risk of dementia compared to people diagnosed in later life,” lead author Christian Schnier, PhD, a senior epidemiologist at the University of Edinburgh, told NeurologyToday.

“It certainly looks like the association of epilepsy with dementia is not cumulative,” he added. He pointed out that due to limitations with data on early life combined with the fact that they didn't have an exact timeline of dementia diagnosis, it was difficult to speculate about the direct relationship regarding time of exposure.

Among the AEDs, exposure to lamotrigine was associated with a 1.5 percent increased risk for AD only. Sodium valproate increased the risk for dementia by 1.6 percent, AD by 1.2 percent, and VaD by 1.7 percent (99% CI 1.3 to 2.3).

Dr. Schnier noted that the increased risk conferred by sodium valproate requires further study. “What we do know is that valproate might have a negative effect on the cardiovascular system through its association with lipid profiles and insulin resistance and the development of the metabolic syndrome. However, not all forms of dementia are related to vascular problems,” he noted.

The authors acknowledged the study's limitations. For instance, the likely variation in how general physician practices transferred medical codes from paper records of patients in early years of life into electronic health records and the fact that some diagnoses could have been misclassified. In addition, the study population was restricted to those followed up from their 60th birthday, and it is uncertain what percentage of study subjects were covered throughout their life.


“It was an interesting and somewhat surprising finding that those diagnosed with epilepsy in early life had a similar risk of dementia compared to people diagnosed in later life,”—DR. CHRISTIAN SCHNIER

Dr. Schnier and his colleagues said the severity and duration of seizure could partially contribute to cognitive impairment. If this is true, improving epilepsy management may decrease dementia risk.

To more thoroughly explore the relationship between epilepsy and dementia, Dr. Schnier said he would conduct a cohort study of dementia risk in people with epilepsy against a group without epilepsy, using follow-up data from birth to old age.

“Further replication studies using different populations/datasets would also help to support our findings and those of others. Given that we suspect an association with cardiovascular health, any research that could show improved cardiovascular health in people with epilepsy would be more than welcome,” he noted.

Expert Commentary

While the nature of the study design limits understanding of causality, there are important take-away messages from the findings of this study, said Sheryl R. Haut, MD, professor of clinical neurology at Albert Einstein College of Medicine and director of adult epilepsy at Montefiore Medical Center.

“The first [lesson] is that clinicians treating patients with epilepsy need to be very sensitive to the possibility of cognitive impairment and/or dementia and be aggressive in screening for this,” she said. “The second is that consideration of medication effects should be made early for any patient with epilepsy, who is exhibiting cognitive issues.”

And lastly, Dr. Haut called upon clinicians to be aware of the increased comorbidity of epilepsy and vascular dementia and suggested that vascular risk factors be treated aggressively in this population.

“The relationship between epilepsy and dementia is complicated. There are many etiologies that can cause both epilepsy and dementia, with stroke being the most common,” said Rani A. Sarkis, MD, MSc, assistant professor of neurology at Harvard Medical School.

He added that unexplained elderly-onset epilepsy has been described as a presentation of neurodegenerative disease.

Given the study limitations—the diagnoses were not confirmed by a neurologist, and there were no details on the episodes of epilepsy or dosages of medications—the authors were correct to label their study as ‘hypothesis-generating,’ he said.

“The take-home message should not be to avoid sodium valproate in the treatment of epilepsy but to devise studies that can determine whether it is a risk factor for dementia,” Dr. Sarkis added.


“The take-home message should not be to avoid sodium valproate in the treatment of epilepsy but to devise studies that can determine whether it is a risk factor for dementia.”—DR. RANI A. SARKIS

In addition to sodium valproate, he noted, the association between lamotrigine and Alzheimer's disease, which is “unexpected, raises the question of confounders such as underlying mood disorders rather than a true causal association” between the medication and the disease.

“Cognitive complaints are one of the most common complaints in the epilepsy clinic, and they should always be taken seriously,” Dr. Sarkis said. “Many issues can cause cognitive impairment in people with epilepsy, including the epilepsy itself, medication side effects, a mood disorder, or a neurodegenerative disease, among other possibilities,” he pointed out. “A cognitive screening battery should be incorporated into the epilepsy visit, especially in people with cognitive complaints.”

“The relationship between an epilepsy diagnosis and the subsequent development of dementia raises several interesting questions,” said David W. Loring, PhD, ABPP, FAAN, professor of neurology and pediatrics at Emory University School of Medicine.

“The finding of increased risk with valproic acid but not with other medications with similar cognitive side-effect profiles argues against any cumulative effect of small but consistent side effects contributing to dementia,” he noted. “Thus, we can speculate that this relationship is related to somatic side-effects of weight gain, with its well-recognized negative health effects including increased vascular disease risk, and medication is a potentially modifiable risk factor for many patients,” he said.

Dr. Loring said that because cognitive comorbidities in epilepsy are often present throughout life, clinicians should consider an appropriate evaluation and diagnosis soon after suspicions are raised about unusual cognitive or behavioral changes. He also emphasized that routine cognitive screening should be considered part of standard care.

“As with other patients with degenerative neurologic conditions, early diagnoses will maximize patient treatment planning and care,” he said. That, he added, will allow clinicians to provide educational material for family and caregivers and encourage patients to engage in activities that “maximally preserve quality of life and independence.”

The study was funded by RS McDonald Trust.


Dr. Schnier reports no relevant disclosures. Dr, Haut is a consultant for Alden Health. Dr. Loring is a consultant for NeuroPace. Dr. Rani has received personal compensation for reading EEGs for Digitrace.

Link Up for More Information

• Schnier C, Dunan S, Wilkinson T, et al. Nation-wide retrospective, data-linkage, cohort study of epilepsy and incident dementia Neurology 2020; Epub 2020 Jul 17.