Continuous intracranial electroencephalography (cEEG) may help pinpoint optimal timing of diagnostic studies and treatment for patients with focal epilepsy, new research suggests.

The findings from a large longitudinal study show that seizure onset in patients with focal epilepsy follows circadian, multiday, and annual cycles.

“Although daily and multiday rhythms have previously been identified, the extent to which these nonrandom rhythms exist in a larger cohort has been unclear,” study investigator Joline Marie Fan, MD, a clinical fellow at the University of California, San Francisco, told Medscape Medical News. “This means that a patient with epilepsy may have a unique combination of seizure rhythms that can inform the days and timing of his or her highest seizure risk,” she added.

The study was published online February 8 in JAMA Neurology.

Distinct Chronotypes

Clinicians and patients alike have long observed cyclical patterns in the onset of epileptic seizures. However, such patterns have rarely been measured in a quantitative way.

Previous studies have examined seizure cycles using inpatient seizure monitoring and patients’ seizure diaries, but the duration of these recordings and their accuracy have been limited. Within the past decade, the advent of cEEG has allowed researchers to observe the cyclical pattern of interictal epileptiform activity (IEA), but the numbers of patients involved in such studies have been limited.

To investigate seizure chronotypes in greater detail, the researchers examined retrospective data for 222 adults with medically refractory focal epilepsy who took part in clinical trials of the NeuroPace responsive neurostimulation (RNS) system.

After implantation in the brain, this system monitors the seizure focus or foci continually and delivers stimulation to stop seizures. Participants also kept seizure diaries and classified their seizures as simple motor, simple other, complex partial, and generalized tonic-clonic.

Fan’s group examined three subpopulations of patients to investigate three durations of seizure cycles. They examined self-reported disabling seizures, electrographic seizures, and IEA. Because patients did not record the time of their disabling seizures, the investigators examined them only in multidien and circannual cycles.

To examine circannual seizure cycles, the investigators included 194 patients who kept continuous seizure diaries for 2 or more years and who reported 24 or more days in which disabling seizures occurred.

To examine multidien seizure cycles, they included 186 participants who reported 24 or more days with disabling seizures during a period of 6 or more months during which the RNS system collected cEEG data. They included 85 patients who had 48 hours or more in which electrographic seizure counts were above zero during 6 or more months of cEEG data collection to examine circadian seizure cycles.

Phase-locking value (PLV) was used to determine the strength of a cycle (ie, the degree of consistency with which seizures occur during certain phases of a cycle). A PLV of 0 represents a uniform distribution of events during various phases of a cycle; a PLV of 1 indicates that all events occur exactly at the same phase of a cycle.

The population’s median age was 35 years, and the sample included approximately equal numbers of men and women. Patients’ focal epilepsies included mesiotemporal (57.2%), frontal (14.0%), neocortical-temporal (9.9%), parietal (4.1%), occipital (1.4%), and multifocal (13.5%). The data included 1118 patient-years of cEEG, 754,108 electrographic seizures, and 313,995 self-reported seizures.

The prevalence of statistically significant circannual seizure cycles in this population was 12%. The prevalence of multidien seizure cycles was 60%, and the prevalence of circadian seizure cycles was 89%. Multidien cycles (mean PLV, 0.34) and circadian cycles (mean PLV, 0.34) were stronger than circannual cycles (mean PLV, 0.17).

Among patients with circannual seizure cycles, there was a weak to moderate tendency for seizures to occur during one of the four seasons. There was no overall trend toward seizure onset in one season among this group.

Among patients with multidien seizure cycles, investigators identified five patterns of IEA fluctuations. One pattern had irregular periodicity, and the others reached peak periodicity at 7, 15, 20, and 30 days. For some patients, one or more periodicities occurred. For most patients, electrographic or self-reported seizures tended to occur on the rising phase of the IEA cycle. IEA increased on days around seizures.

Results showed there were five main seizure peak times among patients with circadian seizure cycles: midnight, 3:00 AM, 9:00 AM, 2:00 PM, and 6:00 PM. These findings corroborate the observations of previous investigations, the researchers note. Hourly IEA peaked during the night, regardless of peak seizure time.

“Although the neurostimulation device offers us a unique opportunity to investigate electrographic seizure activity quantitatively, the generalizability of our study is limited to the patient cohort that we studied,” said Fan. “The study findings are limited to patients with neurostimulation devices used for intractable focal epilepsies.”

The results support patients’ impressions that their seizures occur in a cyclical pattern.

“Ultimately, these findings will be helpful for developing models to aid with seizure forecasting and prediction in order to help reduce the uncertainty of seizure timing for patients with epilepsy,” said Fan.

“Other implications include optimizing the timing for patients to be admitted into the hospital for seizure characterization based on their seizure chronotype, or possibly tailoring a medication regimen in accordance with a patient’s seizure cycles,” she added.

Need for More Research

Commenting on the findings for Medscape Medical News, Tobias Loddenkemper, MD, professor of neurology at Harvard Medical School, Boston, Massachusetts, noted that the study is “one of the largest longitudinal seizure pattern analyses, based on the gold standard of intracranially recorded epileptic seizures.”

The research, he added, extends neurologists’ understanding of seizure patterns over time, expands knowledge about seizure chronotypes, and emphasizes a relationship between IEA and seizures.

The strengths of the study include the recording of seizures with intracranial EEG, its large number of participants, and the long duration of recordings, Loddenkemper said.

However, he said, it is important to note that self-reports are not always reliable. The results may also reflect the influence of potential confounders of seizure patterns, such as seizure triggers, treatment, stimulation, or sleep-wake, circadian, or hormonal cycles, he added.

“In the short term, validation studies, as well as confirmatory studies with less invasive sensors, may be needed,” said Loddenkemper.

“This could potentially include a trial that confirms findings prospectively, utilizing results from video EEG monitoring admissions. In the long term, seizure detection and prediction, as well as interventional chronotherapeutic trials, may be enabled, predicting seizures in individual patients and treating at times of greatest seizure susceptibility.”

The study was supported by grants to some of the authors from the Wyss Center for Bio and Neuroengineering, the Ernest Gallo Foundation, the Swiss National Science Foundation, and the Velux Stiftung. Fan has disclosed no relevant financial relationships.

JAMA Neurol. Published online February 8, 2021. Abstract