Hello, my name is Rana Musavi and I am a neurointensivist at California Pacific Medical Center in San Francisco. Today I'm going to be talking about intracranial pressure and cerebral edema management. I have no disclosures. The learning objectives for today's talk are recognize the clinical manifestations of elevated ICP and cerebral herniation, identify the different types of cerebral edema, describe the physiology of the ICP waveform, and understand the tiered approach to ICP management in patients with acute brain injury. Cerebral edema is defined as the pathological accumulation of excess water in the brain parenchyma. We have different types of cerebral edema. The first is vasogenic, which is an increased permeability of the broad brain barrier, extravasation of proteins, electrolytes, and water. Common causes are brain tumor and cerebral abscess, and it disproportionately affects the white matter. In cytotoxic edema, there's disruption of cell membranes in the brain parenchyma. Water shifts from the extracellular to the intracellular space. Most common cause is ischemic stroke, and it affects both the gray and the white matter. In hydrostatic edema, you have transependymal CSF displacement from the ventricles into the brain parenchyma, and it's mostly caused by obstructive hydrocephalus. Osmotic edema is when the serum osms are less than the brain osms, resulting in fluid movement into the brain, and its causes are hyponatremia, hemodialysis, and DKA. In TBI and intracranial hemorrhage, it's a combination of vasogenic and cytotoxic edema. So the intracranial compartment consists of brain CSF and blood, and as you can tell from this diagram here, about 80% is brain, and then 10% is blood and CSF. The Monroe-Kelly Doctrine basically states that the volume of intracranial vault is fixed, and that these fluids are non-compressible, and any increase in one compartment occurs at the expense of another. So there are compensation mechanisms as intracranial volume increases. As you can tell in the first part of this diagram, we have what's normal. We have brain, the arterial blood, the CSF, and the venous blood, and then we add a mass to this, which causes the CSF and the venous blood flow to decrease. Usually CSF is the lowest pressure compartment and acts as the primary buffer for expanding or space-occupying lesions. However, if that mass is very large, you still see the decrease in the CSF and the venous blood. However, the brain at this point is unable to compensate for the increase in the intracranial pressure. Now we're going to take a look at the intracranial pressure waveform. So this is a single waveform and it has three peaks, and this is what a normal waveform looks like. The first peak, or P1, represents the propagation of arterial pressure through choroid plexus into the ventricles, also called the systolic wave. Then you have P2, which is the arterial pulse further propagating into the brain parenchyma called the tidal wave. And then P3, which represents the closure of the aortic valve called the dicrotic wave. In a brain that has reduced brain compliance, you are going to see that P2 has a higher peak than P1 and P3. P2 can be thought of as a reflection of the arterial pulse that's bouncing off of the brain parenchyma and it represents or reflects brain compliance. Over time, ICP waveforms can take on different morphologies and these are called Lundberg A, B, and C waves. Lundberg A waves are also called plateau waves and they show an increase in ICP for about five to ten minutes. An ICP may rise up to 40 to 50 milligrams of mercury. They reflect a reduced compliance and are always pathological and they may indicate impending herniation. Lundberg B waves show an increase in ICP 0.5 to 2 times per minute. They usually don't exceed 30 milligrams of mercury and indicate poor compliance. And Lundberg C waves show an increase in ICP four to eight times per minute. They usually don't exceed 25 millimeters of mercury. They're seen in normal pathology and can represent interactions between cardiac and respiratory cycles. Cerebral auto-regulation. So cerebral perfusion pressure is equal to the mean arterial pressure minus the intracranial pressure. A normal cerebral perfusion pressure is 50 to 70. Between a map of 50 and 150 is where auto-regulation occurs and you can tell from this graph that the cerebral blood flow remains constant. When the map drops below 50, you have impaired dilation, arterial collapse, and it results in ischemia. And then when the map is greater than 150, you'll have forced mediated dilation, increased flow, and edema. So when the brain goes through hyperperfusion, it causes vasodilation. There's increased cerebral blood volume, which causes an increase in ICP and results in a decrease in CPP. There's a deflection point when compensation mechanisms are exhausted. When cerebral compliance is poor, small changes in intracranial volume result in large changes in ICP. Some physiological changes that can result in increased intracranial volume are flat head positioning, obstructive venous return, hypercarbia, fever, pain, agitation, cough, ventilator dyssynchrony, increased intrathoracic pressure, and increased intra-abdominal pressure. These seemingly benign changes can lead to herniation. Normal ICP is 5 to 15 millimeters of mercury or less than 20 millimeters of mercury. You have to consider the individual. Several small observational studies suggest that controlling ICP less than 20 to 25 millimeters of mercury was associated with improvement in outcomes. The proportion of ICP measurements greater than 20 was most predictive of poor outcomes. There's no large randomized control trials of ICP treatment thresholds. The Brain Trauma Foundation guidelines recommend the treatment of ICP should be initiated when the ICP is greater than 22. Typical manifestations of elevated ICP are headache, depressed level of consciousness, emesis, the Cushing's triad of arterial hypertension, bradycardia, and respiratory irregularity, downward gaze deviation, a sixth nerve palsy, dilated and unreactive pupils seen in uncle herniation, a third nerve palsy also seen in uncle herniation, and a fourth nerve palsy. Some of the clinical manifestations you're going to see are chainstokes, breathing if it's within the cortex, in the midbrain we can see central neurogenic hypoventilation, in the pons we have acoustic respirations, the medulla will result in ataxic respirations, and then anything below that will have respiratory arrest. With uncle herniation and a third nerve palsy you can get a fixed and dilated pupil. In the pons you can get pupils that are mid-position and fixed, and you can also get in the pons pinpoint pupils. If you have upper midbrain damage you can have abnormal flexion, and if it's lower or lower than the red nucleus we'll get abnormal extension. Cerebral herniation is the displacement of brain structures. With subfalcine there's ipsilateral cingulate gyrus movement under the phalx. With central you have downward displacement of the thalami and midbrain and obliteration of the basal cisterns. With upward it usually occurs within the posterior fossa and can happen with the placement of an EVD in that area. Tonsillar you have inferior displacement of the cerebellar tonsils through the foramen magnum and then transcalvarial you have external displacement of the brain through the skull defect. I wanted to focus on uncle slash transcentorial herniation. With this type of herniation you can see compression of the ipsilateral third nerve. Early on you get ipsilateral pupillary dilation. Later on you can get a cranial nerve 3 palsy. You can get hemiparesis contralaterally. There's direct damage to the ipsilateral cortical spinal tract and ipsilaterally you can have Kernhan's notch phenomenon which is contralateral cortical spinal tract compression against the tentorial edge. You can have compression of the cerebral aqueducts which can lead to hydrocephalus. Compression of the ipsilateral PCA which can cause an occipital infarct and then increased pressure on the pons that can lead to direct hemorrhages. As the herniation progresses there'll be extensor posturing and decreased level of consciousness due to midbrain compression and this can lead to central downward herniation and death. Cerebral herniation is not always associated with elevations in ICP. You can have compartmentalized ICP elevations that lead to pressure gradients that displace the brain tissue and global ICP monitoring is not always sensitive. This is particularly the case with posterior fossil lesions that can lead to upward herniation with no changes in ICP. Pupil dilation and uncle herniation usually precedes ICP increases and early recognition of clinical signs and symptoms suggest impending herniation is imperative. So causes of elevated intracranial pressure and or cerebral edema that can be primary intracranial pathology which can be trauma, ICH, subarachnoid, subdural, epidural hematomas, ischemic stroke, tumor infection, hydrocephalus, venous sinus thrombosis, and status epilepticus or they can be from a primary extracranial pathology such as eclampsia, liver failure, DKA, hyponatremia, hypoxia, hypercarbia, high PEEP, hypothermia, obstructed venous return and toxin. So who needs ICP monitoring? In traumatic brain injury, if they have a GCS that's equal or less than eight with an abnormal CT, a GCS that's equal or less than eight with a normal CT, and two of the following, they're over the age of 40, they have signs of posturing or their systolic blood pressure is less than 90. Patients with symptomatic hydrocephalus and patients with non-command following neuroexams and suspected elevated ICP. The mainstay for ICP monitoring is EVD and I'm going to talk about that in the next slide. Some of the complications of an EVD are infection, hemorrhage, shunt obstruction, over drainage of CSF that can lead to a subdural hematoma or misplacement of the EVD. So the different types of ICP monitoring we have and the extraventricular drain, intraparenchymal, subarachnoid slash subdural and epidural. The gold standard is an EVD and the reason for that is that it allows for the treatment and diagnosis. It can monitor the ICP and you can drain the CSF, which allows for the treatment. It has the highest risk of infection, which is about 8% and the highest risk of hemorrhage, which is also about 8%. These are some of the other forms of ICP monitoring. I'm not going to go into those in detail, but you can take a look at them on your own time. Next I want to talk a little bit about the treatment of cerebral edema. So usually with vasogenic edema, we use corticosteroids, osmotic agents, or surgical decompression. With cytotoxic, the treatment options are more limited. You can use osmotic agents as a temporizing measure, surgical decompression in an appropriate clinical context. With hydrostatic edema, the mainstay is CSF diversion, typically with an EVD. And then with osmotic edema, you have to correct the underlying etiology and you can or cannot use osmotic agents. For the treatment of cerebral edema, when it's vasogenic edema, we use corticosteroids. And dexamethasone is usually the mainstay for treatment. And this is usually vasogenic edema secondary to a brain tumor. The role of corticosteroids for vasogenic edema from cerebral abscesses is less clear. The typical dosing is four to six milligrams IVQ6. And it's used to treat the symptoms from the edema rather than the neurological deficit related to the tumor itself. You want to be aware to avoid steroids if you suspect lymphoma. Other than that, you do not give steroids in cytotoxic edema from ICH stroke or TBI. In TBI, they did the crash trial, which showed an increased mortality when those patients were given steroids. And worse outcomes in ICH and stroke have been noted. Okay, so the treatment of intracranial hypertension and hernia, these are the neurocritical care ENLS guidelines. And there are tier zero, tier one, tier two, and tier three. And we're going to talk about each of these in detail. Tier zero represents standard measurements to prevent herniation. This includes elevating the head of the bed 30 to 45 degrees, keeping the head midline, avoiding free water, maintaining normal sodium and serum osmum levels, avoiding hypotension, fever, and shivering, seizure prophylaxis if indicated, control of pain and agitation, avoiding hypoxia and hypercarbia, and rapid sequence intubation if the airway is needed to be protected. Tier one osmotic agents, mannitol and hypertonic saline. They alter the blood dynamics and decrease blood viscosity. Both increase cerebral blood flow and cerebral oxygen delivery, and they extract water from the extracellular space along the osmotic gradient. They're more effective for vasogenic edema, and they need to be tapered off to avoid rebound cerebral edema. Mannitol, reduction ICP is dose dependent. Its effects happen within 10 to 20 minutes with a peak effect between 20 to 60 minutes. Its effects last between four to six hours. Some of the cons is that it can cause renal dysfunction, and in order to prevent that, mannitol administration should be avoided when the serum osmum is greater than 320, but more specifically when the osmolar gap is greater than 20, it may be less effective than hypertonic saline, and you want to make sure that you keep up with the I's and O's of your patient and try to avoid hypovolemia. Hypertonic saline has a rapid onset as quick as five minutes with effects lasting up to 12 hours. Some of the cons are electrolyte imbalances. It can cause central pontine myelinosis, metabolic acidosis, cardiac depression, and congestive heart failure. It can also cause coagulopathies and infusion rate related hypotension. The intravenous administration can cause local effects such as IV infiltration, thrombophilitis, tissue ischemia, and venous thrombus, and there's a risk of AKI if the serum sodium is greater than 160 for a prolonged period of time. Sodium bicarbonate has also been shown to effectively lower ICP. It has the same tonicity as 6% normal saline, and you can give two to three amps emergently. This can be used when it takes too long to get mannitol or hypertonic saline, and you can also consider it for acute seizures or status epilepticus that's secondary to hyponatremia. This is just a graph that compares the osmotic agents. It gives you the loading doses. The route of administration is important because you can give mannitol peripherally, but with hypertonic saline, anything greater than 3%, you need a central line. Both of them, you're going to need to monitor serial BMPs, and with mannitol, you need to monitor the serum osm and hold the dose if it's greater than 320. Some places I've seen them hold it greater than 340 or an osmolar gap of greater than 20. Osmolar gap is the difference between the osmolarity and the osmolality, which is used to detect the presence of unmeasured osmoles such as mannitol, and we talked about some of the risks in the previous slides to each. So which one do I use first? Meta-analysis showed that hypertonic saline was more effective in lowering ICP, but clinical outcomes were not examined. The choice usually depends on IV access and patient factors or comorbidities such as ejection fraction, volume status, and kidney function. The best way to use hypertonic saline is bolus dosing. It's found to be more effective as opposed to a continuous infusion. With continuous infusions, they saw that the brain adapted to higher tonicities and it minimized its benefits. Additionally, there are studies demonstrating that continuous infusions of hypertonic saline for TBI resulted in higher mortality. Therefore, bolus dosing is preferred. Some of the guidelines for osmotic agents and stroke come from the American Heart Association, and they show that osmotic therapy is reasonable for patients with stroke and clinical deterioration due to cerebral edema, but routine osmotic therapy in patients without clinical deterioration is not indicated. So what about a sodium target? So cerium-sodium concentrations do not correlate with ICP values. Remember, it's the creation of the osmotic gradient that causes a reduction in ICP. There are some specific clinical situations that you might encounter when determining which osmotic therapy to use. So if there's no central line, you can use 3% hypertonic or mannitol. If the patient has heart failure, mannitol is preferred. For AKI, 3% hypertonic or 23% if there's a central line. If the patient already has hypernatremia, mannitol. If there's acute hypotension on high-dose steroids, then 3% hypertonic without a central line or 23% with. End-stage renal disease is hypertonic, but it is okay to use one dose of mannitol. If the patient's on the floor and herniating, you can try the 3 amps of bicarb and plan dosing greater than 24 hours. You can do mannitol or 23%. You can also use osmotic therapy when there's evidence of intracranial hypertension and subarachnoid hemorrhage, TBI, ischemic stroke, and intracranial hemorrhage. These are the 2020 Neurocritical Care Society guidelines. For subarachnoid hemorrhage, they suggest symptom-based bolusing of the hypertonic saline rather than target-based, i.e. sodium goals. There's no data that targeting a specific sodium level is beneficial. For TBI, they suggest hypertonic saline over mannitol with mannitol as a second line. And ischemic stroke, you can use hypertonic or mannitol. They suggest against the use of prophylactic mannitol due to the increased potential for harm. For ICH, they suggest hypertonic over mannitol. The literature suggests that hypertonic is at least as safe and effective as mannitol, and the panel felt potential for cerebral perfusion justified suggestion of hypertonic over mannitol. A couple of things I just wanted to reiterate that the osmolar gap correlates better with mannitol concentration. They have no recommendation for osmolar gap cutoff, but the panel recognized that 20 was most often used. And for hypertonic saline, the upper serum sodium limit of 155 to 160 and chloride 110 to 115 was reasonable to reduce the risk of AKI. Other Tier 1 considerations are temporary hyperventilation, a brief course less than two hours of hyperventilation can be considered, a PCO2 goal of 30 to 35. Hyperventilation causes vasoconstriction. There is always the risk of cerebral ischemia. CSF drainage should be considered with acute obstructive hydrocephalus with emergent EBD management. You drain about 5 to 10 cc's of CSF per hour for acute ICP elevations. Another Tier 1 consideration is surgical decompression. If the ICP is not controlled and or clinical signs of herniation do not resolve with other Tier 1 interventions, surgical decompression is an option and should be considered. If surgery is not appropriate, you move on to Tier 2. For Tier 2, you want to maximize the hypertonic saline dose. A sodium greater than 160 was deemed to be unlikely to be beneficial. If the ICP stabilizes, you can maintain the sodium at its current level until the edema improves. You can also increase sedation. Increasing sedation will decrease cerebral metabolic demand and cerebral blood volume and therefore it will decrease ICP. You must manage hypotension associated with any of these sedative agents. It can reduce CPP. In regards to propofol, you must monitor for propofol infusion syndrome which includes metabolic acidosis, cardiac dysfunction, rhabdomyolysis, hypertriglyceridemia, and it can be often fatal. For Tier 3, you can use sedation to titrate to ICP or burst suppression on continuous EEG. Again, you decrease the metabolic demand which decreases the cerebral blood flow and decreases the cerebral blood volume and leads to decreased ICP. We usually use pentobarbital. It carries the risks of hypotension, infection, and paralytic ileus. The common EEG goal is burst suppression 50 to 80% of the time. You can administer this for 24 to 96 hours. It can mimic brain death which includes unreactive pupils. It has a long half-life, 15 to 50 hours, and must be tapered off to prevent rebound ICP elevation. Other Tier 3 approaches are to try hyperventilation to achieve a PCO2 goal of 25 to 35. You can consider this in patients who have failed the other managements. Prolonged hyperventilation is unlikely to be beneficial and may exacerbate ischemic injury. Moderate hypothermia has also been used with a target goal of 33 to 34 degrees. It can reduce ICP but not shown to improve outcomes. You can use external surface cooling, cold saline, or invasive cooling devices. Risks are coagulopathy, shivering, electrolyte disturbances, arrhythmias, and sepsis. One of the trials found that early hypothermia and severe TBI resulted in fewer ICP interventions but overall neurological outcomes were worse. When you rewarm, you must do it slowly to avoid rapid hyperkalemia and distributive shock. I quickly want to mention management in posterior fossa lesions. For example, you have a cerebellar stroke that causes cytotoxic and or vasogenic edema. This can cause compression of the fourth ventricle and direct brainstem compression and a result in obstructive hydrocephalus. The AHA guidelines recommend against EVD placement as primary intervention for obstructive hydrocephalus caused by cerebellar strokes. Consider decompression prior to EVD placement. Intracranial hypertension can also occur in the absence of a large mass lesion. For example, a subarachnoid hemorrhage, bacterial and non-bacterial meningitis, cerebral venous sinus thrombosis. The rate of intracranial hypertension in an aneurysmal subarachnoid hemorrhage is 71%. Higher rates are seen with higher Hunt and Hess and Fisher scores, likely the cause of prehospital mortality. There's a sudden increase in ICP, which leads to a sudden decrease in CPP, which then leads to cerebral ischemia. Why does it matter? Patients with an ICP greater than 20 have a higher mortality rate. Another question is, do you need an EVD before you treat elevated ICP? The BESTRIP trial was a South American TBI trial with patients with GCS 3 to 8, but excluded GCS 3 with bilateral bone or fixed pupil. They were placed into two groups, the intervention group where ICP monitoring and treatment was done to keep ICP less than 20, or the control group that treatment was based on clinical slash imaging studies. The outcome was that there was no difference in survival, duration of coma, and neuropsych status at six months. So with that, does that mean that we don't need to treat elevated ICP? No, it just means that the management guided by a number is not superior to the management guided by CT or exam. Finally, this is a decision tree that can help with the management of these types of patients. If you would first observe a clinical decline in your patient and order a STAT CT head, then you should ask yourself, is there a moderate or high likelihood that that decline was due to herniation, lateral brain shifts, or elevated ICP? If no, you want to determine the underlying etiology. If yes, you can try to give empiric hypertonic or mannitol before leaving the CT or once the patient gets back to the room. And then once you get the imaging, if there is no diagnosis of herniation or brain shifts on the imaging, you again want to try to figure out what the underlying etiology is, and it could be something such as a seizure. But if there is evidence on imaging, then you can determine whether that patient needs intervention, such as a decompressive crani or EVD placement. If those things end up working and the patient does not have any more signs and symptoms of elevated intracranial pressure or they have an EVD that's measuring less than 20 millimeters of mercury, you probably don't need to continue hypertonic and mannitol. If the underlying problem has not resolved and there was no surgical intervention that was deemed appropriate, then you have to develop a plan to continue hypertonic and mannitol therapy and determine what your clinical endpoints would be. I would like to thank Dr. Krista Swisher, who is my mentor and former fellowship director for inviting me to do this talk, and I would also like to thank Dr. Kadena and Dr. Sarwal for allowing me to participate in this course.

intracranial pressure

cerebral edema

ICP waveform

cerebral auto-regulation

elevated ICP

ICP monitoring