Hello, my name is Anthony DiGiorgio, and I will be leading this case-based discussion on the role of ICP in multimodality monitoring in modern TBI. Here are my conflicts of interest. So let's start with a case. This is from our trauma center. It was an unfortunate story of a 25-year-old female who was struck by a car. She was a GCS E3V4 MS6 on arrival. She was confused. She would follow commands in all extremities. She did not recall the incident, was not able to articulate much of a story, and would repeatedly fall asleep when we stopped questioning her. She had obvious signs of trauma, including abrasions and a subcalic hematoma on her posterior scalp, indicating that she had struck the back of her head. This picture here is the classic coup-contra-coup image. You can see the posterior scalp, subcalic hematoma. You can see she has bilateral frontal contusions, temporal contusions. She has a little bit of a tentorial subdural hematoma, and she has overall a relatively tight-looking brain. Because she was able to talk to us, she had a reliable, repeatable neurologic exam. We would, of course, put her in our ICU, but we would not jump to put intracranial monitors in her chest yet. We would give Q1-hour neurochecks, start her on some antiepileptics, which are the standard in our institution. Because of the Q1-hour neurochecks, we were easily able to pick up on this decline in her neurologic exam. Her GCS went from a 13 to 8. We used a pupillometer to get pupillary indices, and those went from 4 to 2, and her heart rate started to decline. These are obviously concerning signs. It sounds like she is about to herniate. At this point, of course, we would consider putting in intracranial monitors. It's also somebody who probably needs airway protection, and we would certainly look to intubate her. So before I talk about what intracranial monitors we would use and how to interpret those results, just want to give an overview of what things we track are in the Neurocritical Care Unit. So we track function. We give the neurologic exam hourly. We can use electrophysiology to track function, either surface EEG, ECOG strips, or even depth electrodes. We track pressures, a number of different pressures, not just the intracranial pressure, the systolic blood pressure, and the cerebral perfusion pressure, and those are all reflective of the Monroe-Kelley Doctrine, which we've depicted here in a little cartoon. We also track cerebral blood flow using a variety of monitors. We can track oxygenation, local oxygenation in the brain using the LICOX, which gives us the brain tissue oxygenation, or the global cerebral oxygenation using a jugular venous oxygen measurement. We can also measure the local metabolism using microdialysis catheters. There are all these different methods that we have managing the cerebral function and the cerebral tissue health. So why do we monitor intracranial pressure? It's one of the oldest biologic parameters we've had to monitor in the ICU. This paper is out of 1981, where they were able to monitor intracranial pressure and found that patients that had an elevated and not reducible ICP had a much higher mortality rate. Evidence here, 92% mortality in patients with high ICPs that were not reducible. Here's a more recent paper from 2011, which shows outcomes based on intracranial pressure. And breaking down this graph, looking at patients with ICPs sustained over 20 and ICPs sustained over 30, they had much worse functional outcomes here and here. So the question is not whether or not ICP is correlated with the bad outcome, but does monitoring ICP affect those outcomes at all? This is the BEST-TRIP trial, which was done out of South America. This famous trial showed that when comparing two different methods of monitoring patients, there was no difference in outcomes. This paper should not be interpreted as the patient should not be monitored. It's that invasive intracranial pressure monitoring was equivalent to extremely close neurologic exam and routine use of serial imaging in these patients. These two papers reflect real-world practice, they were done out of the United States, and show general overall benefit for ICP monitoring. This first paper showed that hospitals with low use of ICP monitors had higher mortality, and the second paper showed that patients who had ICP monitors themselves had lower mortality. This, of course, is subjective to an intention to treat bias. These are retrospective papers, and they're not randomized. However, I think they show good reflection of real-world practice, and that ICP monitoring is a useful surrogate. So what do the guidelines say? There are two guidelines that we look at, the Brain Trauma Foundation and the American College of Surgeons TQIP guidelines. Both of them recommend intracranial pressure monitoring. The PTF guidelines say that monitoring can be used to reduce mortality and severe TBI, by which they mean a GCS of less than 8. The ACS TQIP monitoring continues to have the more rigid prescription for monitoring in patients with a GCS less than 8 and a structural brain injury on CT, patients at high risk of progression, such as those with anticoagulation, and patients who require extracranial surgery for concomitant injuries. Patients like these would not be able to have a close neurologic exam and would be unable to get quick serial imaging if they are in an extended surgery. And there's a few papers showing that there is marked reduction in mortality when these guidelines are followed. Intracranial pressure alone is not the entire story, however. We need to look beyond just simple intracranial pressure and look at other parameters of cerebral health and cerebral perfusion. The cerebral perfusion pressure, of course, is the mean arterial pressure minus the intracranial pressure. It shows how much of the blood flow is actually getting through the edematous brain to provide perfusion. This paper from 1990 showed that with CPP-directed therapy, the overall mortality was much reduced over historic reports. This is a very interesting study from 2005 examining two different institutions and whether or not they used intracranial pressure-directed therapy or cerebral perfusion pressure-directed therapy. In Edinburgh, they used cerebral perfusion pressure, and in Uppsala, they used intracranial pressure. You can see the demographics of the patients here, as well as their treatment paradigms. The interesting part is that these two institutions found very different outcomes, whether or not they used the cerebral perfusion pressure. Uppsala found that cerebral perfusion pressure less than 60 was prognostic of a good outcome, whereas Edinburgh showed that cerebral perfusion pressure less than 60 was associated with death. So these two institutions had completely opposite results. You can see how well the ICPs correlated with the maps in these two different institutions. Edinburgh tended to have ICPs that correlated more closely with mean arterial pressure, whereas Uppsala had less of a robust correlation. So if we then map from these two institutions the probability of favorable outcomes on the y-axis and the map versus ICP slope, the correlation of these two on the x-axis, we find that these two institutions were actually treating very different subsets of patient populations. One of which was autoregulating or had a pressure-active patient population. The other was not autoregulating and had a pressure-passive population, whereby the mean arterial pressure would be transmitted to the ICP, where higher arterial pressure would increase ICP and worsen their outcomes. The other patient population, when the arterial pressure would go up, would not subsequently increase ICP, therefore not worsening outcomes. If we plot these two curves against each other, they clearly show the patient populations that are pressure-active and those that are pressure-passive. This simply reflects the fact that different patients go in and out of autoregulation. And that when autoregulation is intact or is not intact, that should change whether or not we target a transfusion pressure or an intracranial pressure. So what is cerebral autoregulation? It's the ability of the cerebral vasculature to change in caliber and keep intracranial pressure regulated in response to changes in arterial blood pressure. You can see here the cerebral pressure curve compared to the cerebral blood flow. At a very low arterial blood pressure, the blood vessels collapse and there's a linear relationship between arterial blood pressure and cerebral blood flow. However, once blood pressure enters the zone of normal autoregulation, the vessels are able to accommodate changes in arterial blood pressure and keep cerebral blood flow at a constant rate, thereby keeping intracranial pressure at a constant rate. When the body, again, exits autoregulation, the blood vessels become passive to arterial pressure once again. And there's, once again, a linear correlation between cerebral blood flow and mean arterial pressure. This is a picture that I put in showing a brain in surgery that is out of autoregulation. You can see how maximally dilated these blood vessels are. They will not be able to constrict and they will not be able to control the flow of blood through them. So how does brain tissue oxygen play into this whole picture? Well, it turns out that brain tissue oxygen is linearly related to cerebral blood flow and therefore can be used as a surrogate for cerebral blood flow. And we can see, as we map the brain tissue oxygen against the mean arterial pressure, we get a very familiar looking curve similar to this one that we see describing autoregulation. So this is real world application showing this autoregulation curve in real life. So how do we assess autoregulation? Well, we need to look at changes in the mean arterial pressure and see if the cerebral vascular resistance changes. We have multiple ways of assessing the cerebral vascular resistance, but we're simply looking for if these patients are in a realm of intact autoregulation or impaired autoregulation. The way we do it in our institution, we use what's called the pressure reactivity index. It's a linear correlation of intracranial pressure and arterial blood pressure over time that allows for continuous monitoring of autoregulation. It uses intracranial pressure as a surrogate for cerebral blood flow because as the cerebral blood flow changes, the intracranial pressure will also change in a linear pattern. So you can see here the pressure reactivity index 0.95 would indicate there is a high correlation of blood pressure to ICP, whereas in this bottom graph here, a PRX that is negative or flat shows that there's very little correlation between arterial blood pressure and ICP. This top patient is not autoregulating, and the bottom patient is. So here's a good paper that gives an overview of different ways by which we can measure cerebral autoregulation. There's many different indices, but as I mentioned, we tend to prefer the PRX, the pressure reactivity index. It's easy to assess using standard ICU equipment, such as arterial blood pressure and the intracranial pressure monitor. Our method at our hospital is we have patients with A-lines and intracranial pressure monitors. We hook both of those up into the bedside device. It gives us a minute-by-minute output showing how well these two numbers correlate. We can then look back at these graphs and see if ICP goes up in a linear fashion with the arterial blood pressure. If they don't, we can assume the patient is not autoregulation. We'll also check it once a day. We'll give the patient Ebola suppressor under direct observation, and we'll watch to see as the arterial blood pressure climbs, does the ICP climb and correlate with it. If they go up together, we consider this patient to be not autoregulating. If they do not go up together, we'll say the patient is autoregulating. Whether or not the patient is autoregulating will determine our treatment strategy, targeting a CPP goal or an ICP goal. Patients who are autoregulating, we will primarily treat for cerebral perfusion pressure above 60. For patients that are not autoregulating, we'll primarily treat an ICP, trying to keep the ICP less than 20. Now let's go back to our case. We put some intracranial pressure monitors in our patient. We gave her some ventilation, drained off some CSF, and she now has stable ICPs for the last 24 hours. They've been controlled with intermittent CSF drainage. We control sedation. Her MAP and hyperventilation challenge showed that she has intact autoregulation. Now her ICP starts creeping back up. It's 34, and we feel like we have to intervene at this point. So what do we care about? Of course, we want to check the latest labs, ABG and the BMP. You can see here her NPIs are also decreasing. So what are our options? We have many options, many ways to troubleshoot and go from here. But of course, we want to have a standardized framework in how we treat patients. So it's nice to have a succinct protocol. This is the CIVIC guideline. I'm not going to get too much into it, but it is nice to look at because it does take into account some of the more complex intracranial monitors that we've been talking about. It certainly takes into account whether or not the patient is autoregulating. CSF drainage is a simple and straightforward way to treat elevated intracranial pressure. All of our patients get external ventricular drains in addition to an interparenchial ICP monitor. Using the drain allows us to readily drain CSF and provides a quick treatment for intracranial pressure elevations. We use what's called a pressure equalization ratio to show how well CSF drainage changes the intracranial pressure. And you can see here there's a formula that takes into account what the expected ICP change should be based on how much CSF was drained, what the ICP is before and after, and the height of the EVD. You can see in TBI versus non-TBI, there's a general difference in pressure equalization ratios. In non-TBI patients, patients with a purely hydrocephalus problem, the pressure equalization ratio tends to be a lot higher. In TBI, because the problem is likely cerebral edema, the pressure equalization ratio is much lower. Hyperosmolar therapy, of course, is always an option. Manitol versus 23%. We tend to like 23% our institution. It doesn't have as severe side effects with hypotension. And there are some papers showing that it has equivalent efficacy. And of course, we always need to troubleshoot our ICP monitors. If we have a brain tissue oxygenation monitor, we'll give an FIO2 challenge, whereby we increase the FIO2 to 1 and make sure that the PBT02 increases as well. These are very frail monitors, and so they can often get damaged when they're inserted. Or if they're sitting within a blood clot, they will not give a good reflection of the local cerebral blood oxygenation. For the ICP monitors, we want to make sure that our BOLT and our EVDs correlate. We want to check the waveform and make sure it's reflective of reality. If they have a craniectomy defect, we'll actually press on the defect and make sure the scalp pressure transmits to the intracranial pressure. And I always like to check the fluid column on the actual extraventricular drainage tubing, hold it up, make sure that the fluid column is bouncing appropriately, and that it settles at the appropriate level that correlates with the level on the monitor. So in this case, I think it was pretty obvious by some of the lab values. She was a little hypercarbic. By breathing oxygen at PaCO2, we were able to get her ICP down. As we move on to our next case, I'd like to show a picture showing our multimodality monitoring in situ. Here's a CT scan clearly showing there are multiple probes in the brain. As I mentioned before, we put EVDs in all our patients. We also put an intracranial pressure monitor. We will also put an intracranial blood, excuse me, brain tissue oxygenation monitor. We'll put a intracranial depth electrode to monitor for seizure activity deep. In addition, if a patient has a craniectomy or craniotomy, we'll often leave a surface ECON electrode as well. These give a broad picture of overall cerebral health. We'll often use all these numbers to help with clinical decision making. So here's our second case, 42-year-old female, gunshot wound to the head. She's CS12 at the scene, and she declined to a six by the time she was brought to the emergency department. Her CT showed diffused swelling, and her wound was debrided. Should we put in monitoring? Here's a picture of her CT scan. You can clearly see she has a very tight brain. She has a large amount of injury. I think in this case, intracranial monitoring is certainly appropriate. So in this case, ICP is controlled with intermittent drainage and sedation for about 48 hours. And of course, her ICP starts creeping up. Again, we have to think of what are we going to do in this situation. Give her some mannitol, no response. Hypertonic saline, no response. ICP is still elevated. Let's check her labs. PBTO2 is 26. Her ICP is 37. Osteosome is 322. Sodium is high. Her mean arterial pressure is also extremely high. We do an autoregulation challenge, the reverse, because she already has high arterial pressure. We don't want to bump her arterial pressure up any higher. So we can give her a trial of CAR-D. And sure enough, by decreasing arterial pressure, her ICP comes down. This is a patient that's clearly not autoregulating. And in this case, we don't want to push for a certain fusion pressure. We want to push for a lower ICP. Here's one last case presentation. 20-year-old skateboarder, came in GCS7. Here's a CT scan. We get a lot of skateboarders in San Francisco. They love going down the big hills. This is a non-uncommon occurrence. Enter one of the craniectomy. It's just signaling cerebral contusions. We continue to monitor after our craniectomies. They tell us the patient's doing well. Here it is at 2 a.m. Here it is at 4 a.m. Here it is at 6 a.m. This is a few days after surgery. Something to be concerned about is that he is requiring more and more CSF drainage. And while his ICPs are well-controlled, his brain tissue oxygenation monitor is going down. This patient also happened to have a jugular venous oxygen monitor. Resultion was showing a decrease in oxygenation. This is a case where the patient was actually having respiratory problems. He was developing ARGS. And a quick look at his APG and chest X-ray showed that he was not oxygenating well. And by increasing his oxygenation, we're able to stave off a disaster here. So in conclusion, we didn't make sure and realize that ICP is not the only story. Cerebral autoregulation is critical. We need to check that daily to make sure that patients aren't going in and out of autoregulation. If you have a patient where the arterial blood pressure and mean arterial, excuse me, the arterial blood pressure and intracranial pressure are not the same, it's important that patient is not autoregulating and a CPP goal should not be pursued. If the pressure do not go up the same, we should actually favor a CPP goal over an ICP goal. We've had patients where by increasing their cerebral perfusion, you can see a decrease in intracranial pressure. Similarly, PBTO2 monitoring is useful in cases like the last one to alert us to possible cerebral ischemia, worsening cerebral edema due to poor oxygenation. PBTO2 monitoring can also be useful in those cases whereby the patient's going in and out of autoregulation rapidly or whereby patients are on that tenuous cusp where you need to monitor them almost minute by minute. Additionally, we find EEG monitoring very helpful, both with depth electrodes and with surface ECoG. However, all these numbers are extremely difficult to manage as I've tried to show some examples here. Eventually, we will have a need for clinical decision support systems, lead to have data integrated into sort of this cockpit-like scenario whereby clinicians get instant feedback from these machine algorithms. So we're not calculating multiple different variables that are really too much for one person to comprehend at one time. It'll give us actionable information from this raw data that we can then interpret in real time. I think this is the ultimate goal and this is what we'll eventually strive for to help our patients. Thank you for listening today. It's been a pleasure giving this lecture.

intracranial pressure

multimodality monitoring

traumatic brain injury

neurochecks

cerebral perfusion pressure

autoregulation