NeuroVoices: Orrin Devinsky, MD, on Examining the Relationship Between Convulsive Events and Unexplained Sudden Death in Children
The director of NYU Langone’s Comprehensive Epilepsy Center provided perspective on recently published data evaluating the potential mechanisms of sudden deaths in toddlers.
Orrin Devinsky, MD
In 2021, more than 2900 children younger than 4 years died suddenly from unknown causes representing more than 219,000 life years lost. Research on sudden unexplained deaths in childhood (SUDCs) among infants and young children has been limited by the fact that almost all cases are sleep-related and unwitnessed. Published in Neurology, a recent study analyzed terminal videos of sleep-related child deaths from the SUDC Registry and Research Collaborative to document observations and explore mechanisms of death.
Led by Orrin Devinsky, MD, the analysis comprised of 7 consecutively enrolled cases with home video recordings of the child’s last sleep period that were independently reviewed by 8 physicians for video quality, movement, and sound. Results showed that in all 5 continuously recorded videos, a terminal convulsive event lasting 8-50 seconds was identified shortly before death. Abnormal movements were identified in 6 videos and abnormal sounds were found in 4 of the 5 audio-recorded events. Notably, primary cardiac arrhythmias were not supported, as all 7 children had normal cardiac pathology and whole-exome sequencing showed no known cardiac disease variants.
Devinsky, director of NYU Langone’s Comprehensive Epilepsy Center, sat down with NeurologyLive as part of a new iteration of NeuroVoices to discuss the research and how it impacts the neurology community. Also serving as a professor of neurology, Devinsky gave insight on how these data advance the understanding of SUDC and sudden death in epilepsy, as well as how they might contribute to future research. Additionally, he spoke on the ways to potentially prevent SUDC, the advantages of crib cams, and the importance of not over-worrying parents.
NeurologyLive: Coming into the study, how did you want to assess these mechanisms of sudden death in children?
Orrin Devinsky, MD: I'd say in part, we were fortunate or fortunate to get these tragic videos, something we hoped I would never see in my life. But since we know these cases occur, our challenge on the science side is what's causing these deaths, very similar to SUDEP. And I think for SUDEP, we're very tragically fortunate that some of those deaths get captured in Epilepsy Monitoring units. And that was analyzed by colleague and friend, Philippe Ryvlin about a decade ago. And it turns out again, for the patients with epilepsy who are in epilepsy units, the vast majority have a major convulsive seizure, and then their brain kind of shuts off, their breathing gets diminished and then they go into a cardiopulmonary arrest and die. Not that that's every SUDEP, but certainly that appears to be a common pattern. And sure enough, their biggest risk factor are their uncontrolled seizures. When we turn to the SUDC (sudden death in children) population, I personally was always suspicious that seizures might be related to a minority of cases, because a roughly 30% of these children have a history of febrile seizures, which is roughly 10-fold more than the general pediatric population for their age.
There's always a suspicion that some of the deaths might be seizure related. In collecting this SUDC registry that I've done with Laura Gould, MSc, we've gotten about 300 cases. With the advent of technology getting better, many parents have had crib cams, or nanny cams, and just leave them. And then sadly, at some point, their 18-month-old or 30-month-old child dies in the morning, and they find them dead. They go back and look at the camera to see, and they've enrolled them in our registry. We now asked if there is any video footage of the night that the child or time the child passed away. Seven families had footage, and so we analyzed. It's a little bit like being a detective. There's a bank robbery and happened in the middle of the night. No one witnesses were there to say it was this person, they look like that, but we find out there's a camera on the side of the road that looks over the bank, and we look at the videos. I must say I was shocked that of the with good video quality that was continuous recording, all of them showed would appear to be seizures, as independently reviewed by a number of different physicians, including myself.
As a neurologist, what stands out the most among the findings?
I think what concerns me and perhaps stands out the most is that of the seven children, five with good video quality, they all showed what looked to be seizures shortly before death. The other two was really hard to tell, much there looks like a little movement, but just under the covers made it too hard. The other one probably was a convulsion, but just less clear data. Only one of those seven children had a history of febrile seizures. What's remarkable is that no one would have ever suspected a seizure. Firstly, seizures don't leave a mark, especially in this age group. In an older adult, there might be a tongue bite or something like that to be a clue. But in this population, they don't have teeth that are going to cause that kind of injury, and so there's really no marker, no sequelae that you could see. That's the big challenge in this case in these cases, we're dealing with what are probably more physiologic disorders than anatomic disorders.
And the medical examiner system has evolved from the autopsy, which was an incredibly powerful tool. If you go back 100 years ago, someone died, you know what it was, and they might find out, “Oh my goodness, this person had a pancreatic cancer that spread to their liver and their lungs.” But that was never diagnosed in life. The autopsy was very powerful. Someone had an aneurysm in their brain, and it ruptured, and there was a brain bleed, but nobody knew that. It was only determined at autopsy. The whole medical examiner system, autopsy system, has been focused almost exclusively on structural physical abnormalities. That got extended a bit over time with laboratory tests. Now, if someone is found dead next to methamphetamine, we can get a blood level of methamphetamine. And if their levels are high, then the medical examiner will say they likely died from a methamphetamine overdose.
We've now extended the physical world, from organs and tissues and microscopic slides, to toxicology reports and things like that. But we still don't do a great job at measuring physiologic disorders. In America, I would say we're very biased towards heart disease. If a 22-year-old or a seven-year-old dies and there's no cause of death, it will often get listed as cardiac arrest just because it's what everyone just assumes, is the safest bet. Anyone my age, if you look at our coronary arteries, there's no way they're going to be crystal clear. No one over 60 [years old] in America has perfect, gorgeous coronary arteries that look like a 15-year-olds. You’re always going to find something in the heart, it's just a place where you'll see something. In the same way, if you look at the lungs in New York City in the winter, everyone's probably got a little bit of bronchial inflammation. There's hardly anybody who goes to winter without having a cold or an upper respiratory tract infection. That's again where the medical examiner's have often focused. I think one of the important lessons for me is that we need to focus on not what we've been trained and what everyone else focuses on. But since that approach has not given us the answer, it's really important to focus on other areas, and in this case, for physiologic disorders. I do think the shocking lesson is that seizures may be a much more common cause of death in this age group than we suspected. It raises the bigger question, could it be a bigger cause of death in the infant group for SIDS (sudden infant death syndrome)? And the answer is, I think it probably is, to some degree. To what degree, I don't know.
How can we continue to innovate the ways we research sudden death in children?
I think we can. For example, these kinds of crib cams are relatively inexpensive. The problem is, we have no idea which child is going to suffer some one-in-5000 tragedy. But I do think by collecting more data from those parents who do have these video recordings will be very important for us to understand was just a weird chance event that five out of five, showed this? If we had 100, would it be 20 out of 100? Or would it be 80 out of 100? or more? I don't know. But I think getting that data is really important. And then I think we can turn our attention to the question of prevention. Something like a sock which they make commercially, not terribly expensive, that can give oxygen readings. For most of these mechanisms, whether it's a cardiac cause of death, or neurologic epilepsy cause of death, presumably, there's an event cardiac arrest, a seizure, followed by a period of impaired breathing. If a child had an oxygen monitor, or let's just say a sock that gave an oxygen read, then you'd be able to alarm the parents in the bedroom and say, "your child's oxygen went under 90%, you better get in there and check out what's going on." And that probably would be lifesaving in many cases.
Is there a way to prevent convulsive activity prior to/during sleep?
As I said, only one of the seven children had known history of febrile seizures. Even for the febrile seizure history, think about it, 3% of all children in the United States have febrile seizures, hundreds of thousands every year have febrile seizures. The challenge is that it’s an extraordinarily rare event, and we also don't want to panic 10,000 parents in the hope of potentially saving one life. I think this becomes a challenging issue in public health. My own view is it is important to empower people with information and let them make a choice.
Another metaphor or example is if you went to the supermarket and it's a two mile ride from your apartment, if you said, “Should I wear a seatbelt? Or does it really matter? It's just two miles,” most people would say, “Yeah, you're right, the chance you're going to get into a tragic accident on a two mile ride to the supermarket on a given Sunday morning is extraordinarily low.” But no one would say don't wear the seatbelt, you don't really need it. We're all used to doing things in our lives that have minimal cost: putting a seatbelt on. Maybe it'll crease your shirt, maybe it's a little uncomfortable, but other than that, it's pretty benign. If you didn't have one in your car, it might cost money to get it but now their standard issue, so to speak.
There's not much of a reason to not the wear seatbelt during the two mile ride. In the same way, perhaps we think about very inexpensive ways to monitor children. One of the key is not having too many false alarms. We don't want parents being woken up twice a night, seven days a week, to prevent something that may happen to one in 10,000 children. That's where we have to think about the preventive measures. As we move forward in medicine, one of the transitions is that these decisions shouldn't be on the doctor side alone, the pediatrician or the pediatric neurologist, it shouldn't be a discussion with the parents. If you're a young parent, and I said, "Hey, it’s a tiny, tiny, tiny, tiny risk, but if you wanted to get a sock for your child that read oxygen and it could alarm you and your wife in your bedroom, that's something you could do." The chance it'll ever be needed is extraordinarily small, but it's a preventive measure you could take.
Regarding the abnormal movements/sounds observed, is there anything notable parents should be aware of?
No, I think these were not incredibly loud or obvious. These were things recorded that we had to go back and play it several times and listen. So they're often kind of plaintive, just small sounds, or a vocalization that probably occurred during a seizure. But these were not loud, these would never wake up, and they didn't wake a parent up. They tend to be pretty soft and plaintive, they're not dramatic in the way that one might imagine some seizures in adults, which can be pretty loud. The vocalizations that occur with these children, because the nature of the child and the age, tend to be much, much milder
Where do we go from here? How do we continue to build on this data?
We're trying to be as comprehensive and open-minded and not have a specific hypothesis. We've done whole genome sequencing on all 300 children in the registry, sequencing the 20,000 genes. Those have occasionally revealed something, but in the seven cases, they were perfectly normal. We don't think there was a cardiac cause of death. I think getting more cases, getting more data on more cases, like these videos, is incredibly valuable. Now that this is out in the public, and people are aware of it, I would hope in the future, we are able to get more parents who have near death events, or God forbid deadly events, who would be willing to share those very precious and painful videos with researchers so we can learn more.
Transcript edited for clarity.
