Subscribe to eTOC

Study of US Veterans Points to a Connection Between Late-Onset Seizures and Dementia

Article In Brief

An analysis of data from the US Veterans Health Administration National Patient Care Database suggests that late-onset seizures of unknown etiology are associated with an increased risk of dementia in older veterans.

Late-onset seizures may be an early sign of dementia, according to a retrospective study involving nearly 300,000 veterans.

Veterans who were diagnosed with late-onset seizures of unknown cause were nearly twice as likely to be diagnosed with dementia over the next several years.

“While seizures are commonly thought to occur in late stages of dementia, these findings suggest unexplained seizures in older adults may be a first sign of neurodegenerative disease,” said the authors of the study, reported online March 9 in JAMA Neurology.

Ophir Keret, MD, the study's first author and an Atlantic Fellow at the Global Brain Health Institute at the University of California, San Francisco (UCSF), said that while more research is needed on the relationship between late-onset seizures and dementia, the new findings should “encourage clinicians to follow these patients more closely for cognitive decline and possible intervention.”

The incidence of unprovoked seizures and epilepsy increases later in life, with a two-fold to six-fold higher incidence in older adults compared with younger people, the study authors pointed out.

“These high rates are because of age-related accumulation of seizure-related structural disease of the brain, such as cerebrovascular disease (CVD), intracranial tumors, traumatic brain injury (TBI), previous brain infections, and neurodegenerative disease,” the investigators wrote. But a cause cannot be identified in about one-third of cases of late-onset unprovoked seizures.

That there is some connection between late-onset seizures and dementia is not a new idea. The paper noted that people with dementia are more prone to seizures. Also, mouse models of Alzheimer's disease (AD) suggest that “high rates of seizures are secondary to AD pathology, specifically β-amyloid (Aβ) oligomers.”

Research in such mouse models suggests that amyloid-beta deposition impairs synaptic plasticity, which results in aberrant network excitation leading to epilepsy, the study authors noted.

Those findings plus others indicate that it's “plausible that dementia-associated neuropathology, such as Aβ accumulation, causes seizures in some patients years before they manifest clinical signs of dementia.”

Study Design

Up to now there have been few human studies, however, examining the possible link between late-onset seizures and dementia, and results of those studies have been mixed.

This new study used data from the US Veterans Health Administration (VHA) National Patient Care Database, which contains all inpatient and outpatient encounters within VHA healthcare facilities.

From a sample of 941,524 veterans age 55 or older seen between October 2001 and September 2015, the researchers excluded anyone without follow-up data or whose seizures could be due to neurotoxin exposure (including drug or alcohol abuse), cardiovascular disease, previously diagnosed dementia, brain tumor or brain infection, as well as those who had previously experienced seizures.

That left a cohort of 292,262 veterans, most of them men (96.7 percent), with a mean age of 73. The incidence of late-onset unprovoked seizures was assessed before or during a five-year baseline period. Of the nearly 300,000-person group, 2166 vets (0.7 percent) developed late-onset seizures. They were more likely to be younger, black, have lower income and a higher prevalence of comorbidities at baseline.

During a mean follow-up of 6.1 years after seizure onset, 14,076 veterans (1.1 percent) including 181 vets with late-onset seizures, developed dementia. The vets who had late-onset seizures had roughly two times the risk of developing dementia than veterans who did not have late-onset seizures, the researchers reported.

The elevated risk remained after the researchers did statistical adjustments to account for demographic differences and medical comorbidities. When statistical adjustments were made for depression and traumatic brain injury, including mild and moderate cases, the higher risk of dementia associated with late-onset seizures was attenuated slightly but remained elevated compared to vets who did not have seizures. The risk remained the same when the researchers looked only at cases with at least a two-year time lag between seizure occurrence and diagnosis of dementia.

“While seizures commonly occur in later stages of dementia, our results suggest it can also precede dementia by several years,” the researchers concluded.

“These findings emphasize the need for comprehensive research on late-onset seizures and dementia,” Dr. Keret said. “Considering the growing elderly population and high incidence of late-onset seizures of unknown etiology, further studies may have important ramifications for dementia research and clinical practice.”

Dr. Keret, who conducted the study with senior investigator, Kristine Yaffe, MD, professor of psychiatry, neurology and epidemiology at UCSF, said a prospective study of non-veterans with late-onset seizures of unknown origin would help to better understand the risk and course of cognitive decline.

Because the UCSF study focused solely on veterans, almost all male, the findings may not apply to the general population. People who get their care at the VA also tend to have poorer health overall, which may have influenced the results. Another limitation of the study is that it relied on diagnostic codes, which aren't always accurate, as opposed to reviewing medical records. Data on seizure severity and subtype as well as dementia subtype were limited.

Expert Commentary

Lawrence S. Honig, MD, PhD, FAAN, professor of neurology at Columbia University College of Physicians and Surgeons and Director of the New York State Center of Excellence for Alzheimer's Disease, said he did not believe any definitive conclusions could be drawn from the retrospective study, in large part due to its reliance on a database of clinical diagnoses. He said that while the study might indicate that late-onset unexplained seizures were the first sign of neurodegenerative disease, there could be other explanations as well.

Dr. Honig said the veteran population from which the diagnostic data was drawn overall had a high incidence of morbidity and mortality. For instance, 10.1 percent had diabetes, 17 percent had hypertension, 8.1 percent had depression, and 11.2 percent were obese; 20.1 percent died during the outcomes period.

“That means they are sick people, and the sicker you are, the more likely you are to be diagnosed with dementia and to have seizures,” Dr. Honig said.

“While the study did do statistical adjustments for certain conditions, diagnostic codes alone can't capture the nuances of disease that may or may not be captured in medical records,” he said. He also noted that older persons may experience syncope or fainting that may include movements that result in an incorrect diagnosis of seizures.

Dr. Honig has done research looking at the reverse question, of whether dementia causes seizures. He coauthored a study in 2009 in Archives of Neurology that found that 1.5 percent of 483 Alzheimer's patients followed prospectively for a mean of seven of eight years had unprovoked seizures, higher than the general population but lower than what some experts thought.

Thomas Grabowski, MD, professor of radiology and neurology and director of the University of Washington Alzheimer's Disease Research Center, said he welcomed the new study of veterans because “it contributes to the literature that is raising awareness of a connection between Alzheimer's and epilepsy.”

But he does not think the findings “are enough to be actionable clinically,” in the sense that all patients with late-onset seizures of unknown origins should be put through a detailed workup for dementia or Alzheimer's disease even though there may be an elevated risk.


“Considering the growing elderly population and high incidence of late-onset seizures of unknown etiology, further studies may have important ramifications for dementia research and clinical practice.”—DR. OPHIR KERET


“While the study did do statistical adjustments for certain conditions, diagnostic codes alone cant capture the nuances of disease that may or may not be captured in medical records.”—DR. LAWRENCE S. HONIG

At the same time, Dr. Grabowski said he is quicker to use low-dose anti-convulsant drugs in patients with mild cognitive impairment and episodic symptoms because of EEG studies showing increased electrical excitability in the hippocampus of Alzheimer's patients, and other reports that Alzheimer's patients are at elevated risk for seizures.

Andrew J. Cole, MD, FAAN, professor of neurology at Harvard Medical School and director of the Massachusetts General Hospital Epilepsy Service, said there is reason to think, from both animals and human studies, that “there is a huge story around the relationship between seizures and degenerative disease.”

Dr. Cole said the new study using VA data “provides an interesting directional indicator that certainly would justify prospective study follow-up.”

But he also cautioned against making too much of the study findings. He said he was recently at a medical conference where it was suggested that 50 percent or more of the diagnostic codes may be erroneous, and that can be true for both dementia (especially mild and moderate) and seizures.

Dr. Cole noted the type of dementia was not specified. Also, without medical records, there is no way to know whether in the cases of unprovoked late-onset seizures of unknown origin, there was a thorough workup (using MRI, not just CT, for instance) for pre-existing conditions such as an old stroke or other subtle brain injury that would have excluded them from the study and could independently increase the risk for both seizures and dementia.

Dr. Cole said the elevated dementia risk in patients with seizures identified in the veterans study might also occur because the vets who had seizures had had more intensive neurological scrutiny and follow-up checkups, which would increase the odds of being diagnosed with dementia compared with vets who had not been seen for a neurological condition.

“While there is no doubt a bidirectional connection between seizures and dementia,” Dr. Cole said that “when it comes to the underlying mechanisms, there is a lot of biology to be worked out.”

He noted that EEG studies of patients with early Alzheimer's to look for seizure-like activity are challenging because electrodes placed on the scalp may not pick up electrical excitability deep in the hippocampus. He published a study in 2017 in Nature Medicine that involved placing intracranial electrodes through the foramen ovale that detected silent hippocampal seizures and spikes during sleep. The study raises the hypotheses “that early development of occult hippocampal excitability may contribute to both cognitive symptoms and disease progression in AD.”

Andrew E. Budson, MD, chief of cognitive and behavioral neurology at VA Boston Health Care System and professor of neurology at Boston University School of Medicine, said the study on veterans needs to be replicated before definitive conclusions can be made.

He added: “It makes sense for doctors who evaluate patients with late-onset seizures to ask the patient and family questions related to forgetfulness, memory loss and disorientation that might lead one to consider a diagnosis of dementia.”

He uses the Montreal Cognitive Assessment to screen patients who he suspects may have dementia.

Dr. Budson said detecting cognitive impairment at the earliest stages is important because families need to be aware that a loved one might forget to take their medicine or turn off the stove.

Dr. Budson said he hoped future studies on the connection between late-onset seizures and dementia provide detail on the type of dementia, such as whether the patient has evidence of vascular dementia versus Alzheimer's disease, as well as imaging and biomarker information.


Dr. Budson has received consulting fees from Sage Pharmaceuticals, Cognito, Corium, and research funding from Biogen. Dr. Cole has received consulting fees from Sage Therapeutics, Medtronic (as a chair of the data safety monitoring board), and Vertex Pharmaceuticals. Dr. Keret had no disclosures.

Link Up for More Information

• Keret O, Hoang TD, Xia F, et al. Association of late-onset unprovoked seizures of unknown etiology with the risk of developing dementia in older veterans JAMA Neurol 2020; Epub 2020 March 9.