I don't have any relevant disclosures. I'm going to go through a review of the definition of severe neurology impairment in children. I'll set that in the context of the spectrum of pediatric neurocritical care, highlight some issues arising from a typical recent case in RICU, review briefly the literature on neuroprognostication for cardiac arrest and traumatic brain injury, and talk about where I see the considerable gaps in this field. So just to set the definitions to be clear, so severe neurologic impairment is defined as a group of disorders involving a central nervous system arising in childhood, resulting in motor and cognitive impairment, includes the idea of medical complexity, much assistance being required for the activities of daily living, and the impermanent can be static or progressive. So this was arrived at as a consensus as a Delphi method, and there's some important themes to highlight here. First, the idea of complexity. This reflects the fact that this is a description, not a diagnosis, and that the complexity of these cases is such that it is beyond the knowledge or expertise of a single medical discipline to take care of these children, or indeed to provide neuroprognostication for these families. We all do this, but we all do this representing different disciplines. It's also important to remember this is a consensus statement, and this consensus was largely based on input from English-speaking countries, mostly from a single country, and that within that, there are things that were included late or excluded late, so multi-organ dysfunction, dependence on medical technology, family-reported issues. They didn't make it into this consensus. That doesn't mean they're not important. They just weren't included in this particular definition. This is a good amount of literature for the prediction of mortality with neurologic impairment early in life, and this is a nice literature or analysis of cohort studies performed by Abuga and colleagues, where they looked at studies involving children with childhood onset of epilepsy, intellectual disability, or cerebral palsy, in which there were data predicting outcome, mortality, actually, specifically. And amongst these 2,100 or so studies, 24 met criteria for inclusion in this analysis, and there's a couple of things that are striking about this and relevant to discussion. First, this is a large analysis, but the primary outcome is mortality, which we know is relatively low, at least compared to adults in critically ill children, and that's only part of the questioning that parents pose to us when they're asking what the outcome for their child would be. Within that, there's interesting and important variability. Gender had a variable effect as it aged, important in some studies, not in others. In general, the younger the age, the higher the risk for mortality. Specific to these particular disorders they analyzed, for epilepsy, the standardized mortality ratio compared to healthy children was sevenfold higher, and highest in the first 10 years of the diagnosis. For intellectual disability, also higher, but not as high as epilepsy, but particularly higher in children with cerebral palsy, a standardized mortality ratio above 25, and highest in the relatively young children. Taking these results and applying them to an individual patient is fraught with challenges. One of the reasons for that is reflected here. This is simply a description of what we know to be the case of the spectrum of pediatric neurocritical care. Although this large analysis I just showed you comprised epilepsy, cerebral palsy, intellectual disability, which actually reflect the disorders of many children who have frequent readmissions to the ICU, even if they have those chronic neurologic conditions, they're readmitted with other medical complications. But within the spectrum of children with neurologic morbidity in the ICU, this extends from traumatic brain injury, cardiac arrest, stroke and all its variations, neuromuscular disorders, genetic and metabolic disorders, the epileptic encephalopathies, and then the neurologic complications of sepsis, diabetes, bone marrow transplant, and the myriad of other neurologic morbidities that occur in the ICU. And it's not just those disease labels, it's the mechanisms that go with that, from blood brain barrier dysfunction, ischemia, reperfusion, cerebral autoregulation, neuroinflammatory disorders, glial-mediated injury, neuronal-mediated injury. So now, you have to think about the spectrum of the ability to prognosticate and the limitations of prognostication, not just within the limitations of a label of a particular disorder, but within the limitations of understanding what are the monocular mechanisms of all these complex medical injuries occurring at different developmental stages. This is a recent case from our ICU, developmentally typical, 14-month-old girl, no perinatal risk factors, woke up, had a fever, and woke up from a nap, didn't look right to her parents, brought her to the emergency room, she was staring with no response to visual threat. She was bradycardic, febrile, she had autonomic dysfunction, sluggish and reactive pupils, had transaminitis, and was positive for rhinoenterovirus and HHV-6. So here we have, for the parents, an acute deterioration in their child of uncertain mechanism and uncertain outcome, MRI obtained within the first 24 hours showed restricted diffusion of periventricular white matter, the thalami extending down at the brainstem with multiple microhemorrhages. And the imaging and the course were rapidly consistent with acute necrotizing encephalopathy. This after genetic testing was also sent, and you can imagine an extensive metabolic and other infectious evaluation. Over the course of the next 48 hours or so, she developed refractory status epilepticus, intracranial hypertension requiring an EVD. And over the course of the next couple of weeks, genetic testing revealed a variant of uncertain significance in the SCN2 gene, a sodium channel. So gradually, the diagnosis converged on the acute necrotizing encephalopathy from a combination of HHV-6 encephalitis and possibly an underlying genetic susceptibility. Now, all of these decisions about care, managing of seizures, the management of intracranial hypertension, immune modulation, all those decisions were taking place within the first 72 hours or so. And the parents were asked to participate in the decision-making about delivering that care with understanding the risks and benefits as we understood them at the time. Kerry will talk in a few minutes about the zone of discretion for parents. And part of that zone of discretion for me is exemplified in this case is watching the parents as they transitioned from the ICU to the floor, to rehab, as they became more informed about, as we became more informed about, as they developed the skills to be able to ask more questions, more critical questions about what was in the best interest of their daughter, something that never had the reason to occur to them the day before they came in for their submission. So the evolution of the parents' expertise was a striking part of the development in this case as well. And this was particularly salient four months after recovery, where probably likely due to the thalamic injury, this little girl still does not have a good sleep-wake cycle, her eyes open, she's not responsive, she has dystonic posturing, she's coming through the phase of sympathetic hyperarousal as well, and her EEG just shows mostly slowing. The themes that this case brings out are some areas of uncertainty, the things that we say with rather dispiriting frequency to families, which is the exam will change, there's a spectrum of recovery, we don't know exactly what this will mean for your child, but it will be slow, therapy is important, long-term therapy after your child leaves the ICU. But and this relates to the importance of understanding mechanism. Without an understanding of mechanism, it's very hard to be certain about outcome. Uncertainty about mechanism is a cardinal feature of pediatric neurocritical care. It's also admixed with the brain development, which is even at the older children, is still going on for our patients. And so decisions about interventions must be made in this zone of considerable degree of uncertainty. And for parents who need to have some expert advice and be part of the decision-making process early in a child's course, their ability or the skills that they acquire, the expertise they acquire, changes over time as well. And the other thing a case like this exemplifies, and it goes to the inclusion in the idea of severe neurologic impairment as being something beyond the expertise of a single discipline. The advice to the parents and the involvement in the decision-making for their daughter is based on input from intensivists, palliative care physicians, neurologists, to some extent geneticists, and the rehab doctors, all of which have changed over time. We do have, in this literature, some amount of data on predicting outcome after cardiac arrest and traumatic brain injury as limited as it is. This is a recent analysis by Dykland and colleagues, where they examined studies of multivariable models for prediction of outcome after traumatic brain injury. And this large set of studies, actually there were 54 studies, including 48 models, included children above the age of 14 years, the remainder largely were adults. The striking thing about this is that the outcome measures that were analyzed in this large analysis were really, from a pediatric perspective, relatively crude. Mortality, again, and the extended Glasgow outcome score at a year. And the measures that emerged looking at studies were, there was a model developed and validated at another center, age, the motor Glasgow outcome scale score, and pupil reactivity, and then a myriad of others as well. So even for traumatic brain injury, using this kind of analysis, the ability to predict is based on relatively crude measures. The authors comment in this really well done and interesting paper that any prognostic model should be considered in addition to clinical experience. So that raises the question, then, what counts as clinical experience, particularly in the context of the things that I've been highlighting, that you're all well aware of, of the incredible degree of complexity that's needed to provide useful information for parents based on the complexity of neurologic injury in the developing brain. This is exemplified, I think, in the recent publication on guidelines for prediction of neurologic outcome after cardiac arrest in adults. This is a meticulously analyzed paper using great, great criteria, and the guidelines in detail you can find in blue in the table there and summarized on the right. Must leave time, we use a combination of the exam, imaging, and the EEG. What's not addressed in this paper, and there's a real gap in our literature, I think, is who actually has the expertise to carry out this analysis? Who should be delivering this information? The paper itself is an outstanding analysis of the literature that's available. What counts, what then, what kind of expertise and experience do you need to be the person who is taking this literature summary and using it to apply that information to an individual child? That's simply not addressed, not a flaw of this paper, but a gap in the way we approach neuroprognostication. There was an important paper published with respect to cardiac arrest by Alexis Topchin and colleagues at CHOP just last year. This was only a single-center study, but an important contribution to the prediction of outcome after pediatric cardiac arrest because they looked at outcome after a median of five years after out-of-hospital cardiac arrest in children. So in their single-center study, they had 44 subjects, median age at arrest was five. They set as a criterion for inclusion in this analysis a recovery period of at least a year and the median recovery was five years. One of the important findings in this report, even though it used a long-term follow-up was by phone call, was that when they looked at the association of the pediatric cerebral performance category at discharge, it had a very limited association with long-term outcome. You can see this in the figure here. There are some patients who had a poor CPC, three or four, at the time of discharge from the hospital. In some cases, that improved. In some cases, a good CPC actually worsened. Long-term outcome is really important. We know that. It's a platitude. But predicting long-term outcome, the data we have available in our literature to predict long-term outcome based on neurologic function or exam at the time of discharge is very limited indeed. So I've tried to make the case, I've summarized here, I've emphasized that severe neurologic impairment is a diagnosis, is a description, not a diagnosis. It's a description of a class of disorders that require the input of multiple medical disciplines. Our current literature for prognostication is highly biased towards mortality and short-term outcome. Early stages of recovery, as the CHOP cardiac arrest paper exemplifies, do not predict long-term outcome. That's a reflection of the multiple mechanisms of injury that occur in children with acquired primary or acquired neurologic injury. And when models, if we are able to develop models for neuroprognostication, they need validation and updating, and they need to be calibrated for specific settings for specific centers. So there are different responses to how each medical center practices, which are likely important. And above all, I would argue that this is, neuroprognostication is actually a skill that requires specific training and above all humility. Thank you.

pediatric neurocritical care

neurological impairment

multidisciplinary approach

neuroprognostication

long-term outcomes