Article
NeuroVoices: Orrin Devinsky, MD, Heart Rate Variability as a SUDEP Biomarker
Author(s):
The director of the Comprehensive Epilepsy Center at NYU Langone discussed his recently published research, which identified a novel biomarker for sudden unexpected death in epilepsy.
Orrin Devinsky, MD
For some people living with epilepsy, the risk of sudden unexpected death in epilepsy (SUDEP) is an important concern. Although some forms of refractory or drug-resistant epilepsy may present more risk for SUDEP, clinicians have struggled to calculate the relative risk of SUDEP without a constant biomarker. Recently, in the largest SUDEP biomarker study to date, senior author Orrin Devinsky, MD, and colleagues identified heart rate variability (HRV) as a novel biomarker for SUDEP.
The retrospective, nested, multicenter study featured 31 SUDEP cases and 56 controls. Normalized low-frequency power (LFP) during wakefulness was lower in SUDEP cases (median 42.5 [interquartile range (IQR), 32.6-52.6) than epilepsy controls (55.5 [IQR, 40.7-68.9]; P = .015; critical value = .025). Additionally, there was a negative correlation between LFP and the latency to SUDEP, where each 1% incremental reduction in normalized LFP conferred a 2.7% decrease in the latency to SUDEP (95% CI, 0.95-0.995; P = .017; critical value = .025). Survival models using normalized LFP were found to be associated with SUDEP (c-statistic, 0.66 [95% CI, 0.55-0.77]), which non-significantly increased with the addition of normalized HFP (c-statistic, 0.70 [95% CI, 0.59-0.81]; P = .209).
Devinsky, director of the Comprehensive Epilepsy Center, NYU Langone, and professor of neurology, NYU Grossman School of Medicine, sat down with NeurologyLive to discuss the premise behind the study and why the results carry so much weight clinically. He also provided insight on what has been previously observed in terms of SUDEP biomarkers, and whether these results change the way clinicians conduct clinical trials and the inclusion criteria being used.
NeurologyLive: Can you provide further detail on your results and the clinical significance of them?
Orrin Devinsky, MD: We’ve been struggling in the SUDEP research community to identify a biomarker. The reason it’s important is that SUDEP is so tragic, affecting a minimum of 3000 American lives a year, quite possibly 4000. We don’t know the exact number, but when it happens it can be a health 20-year-old college student, a 40-year-old husband or wife, and it’s just unimaginable for these families. It’s unpredictable. We know some things, if people have frequent tonic-clonic seizures, nocturnal tonic-clonic seizures, they’re going to be at quite a high risk. For most patients with epilepsy, they don’t fit into those categories and are still at risk.
We need to do better. In addition to helping patients understand what their risk is so that we can treat them better, it’s also the ability to assess interventions. If we were to get an intervention, let’s say a watch that detected a seizure and gave out a lout alarm to awaken a post-ictal patient, then we would have to follow so many people for so long. That study would be prohibitively expensive. If only we had a biomarker that identified a patient that looks well controlled but is at high risk. The only biomarker that’s been proposed that has some support, is post-ictal EEG suppression, which certainly makes logical sense. If prolonged, those individuals from some studies, not all, but have been at higher risk for SUDEP.
Having said that, the problem with post-ictal EEG suppression is to get that, you must admit someone to a hospital and take them off their medications. They may be in for a week and not have a seizure and the whole thing may cost $50,000 or more, depending on the city and the environment. Less than 5% of epilepsy patients, my guess is closer to 2% or 3%, in the United States, have PGES. The ideal biomarker is not going to be something that’s prohibitively expensive, puts people at risk, and is extraordinarily hard to get. The EKG is available everywhere, every hospital, most doctors’ offices, outpatient facilities. The cost is phenomenally less than even a routine EEG, but certainly less than an inpatient video EEG. It’s always been an attractive potential area.
We’re not the first ones to go down this road. In cardiology, we know that certain markers of low frequency power, for example, and high frequency power, can help differentiate between high and low risk for sudden cardiac death. In the same way, our most robust finding was that during wakefulness normalized, low frequency power, which is one of the main heart rate variability measures, was lower in SUDEP cases than the epilepsy controls. That correlates exactly with the sudden cardiac death risk. We now have something that can identify a population, relatively inexpensively. To date, this is the largest study looking for biomarkers in SUDEP. Others have been done, but we had 31 SUDEP cases with 56 controls who had epilepsy that were studied at a similar time. We hope that we can get more cases and replicate this finding. I think it already looks like it may well be the most promising, obtainable, SUDEP biomarker.
How do these findings impact clinical trials? Does this change inclusion or exclusion criteria?
I think it will. No one has really gone after a SUDEP prevention study to look at an intervention like a watch because it’s just too many patients for too much time. With this, we could screen for example, 500 patients with epilepsy, find the 80 who are at higher risk, and then randomize them to 2 groups. One will get the watch and one doesn’t, or something like that. There’s ethical concerns and all those things. If you could design a trial where both groups get the education, one group gets the seizure detection watch, and proved that it works, then it will not only be something that insurances will cover much more readily, but it can become a standard of care and allow of a little bit of pressure on not just epilepsy specialists, but neurologists in general, to prescribe this and use it. But we’re not there yet.
Has there been any research on variability of heart rate based on epilepsy subtype?
Not that I’m aware of yet. It’s been looked at. Now that there are more people with longer-term devices, from Fitbits to Apple Watches, we’re going to start to get that data. For now, we don’t have it.
Transcript edited for clarity.
