Chapter 65. Intracranial Pressure: Monitoring and Management

• Intracranial hypertension is the final common pathway of morbidity and mortality for diverse neurologic problems, and its proper treatment requires the timely application of the available therapeutic alternatives when the clinical situation and prognosis warrant and justify treatment.
• Anticipating patients at risk for brain swelling is important and allows the development of a strategic plan to limit its severity and facilitate early intervention when possible.
• The initial therapeutic focus for intracranial pressure reduction should be the control of factors that can aggravate intracranial hypertension, such as inappropriate head and body position, elevated body temperature, inadequate treatment of pain and agitation, elevated airway pressures, blood pressure fluctuations, seizures, and administration of hypotonic fluids.
• The appropriate conventional medical therapies should be selected based on the details of each specific case. It should be clear that brain tissue displacement (BTD) and intracranial hypertension are not synonymous, and the treatment of each of them can be different. Surgical removal of an intracranial mass lesion or expansion of the intracranial compartment should be considered in patients with severe BTD and/or evolving intracranial hypertension.
• In the end, the treatment of intracranial hypertension is heuristic, challenging the managing physician's thorough understanding of the cause of the problem and his or her ability to define the human aspects that relate to determining the appropriate level of care for individual patients. In addition, successful management depends on a partnership with nurses founded on comprehensive communication to assist with the accurate translation of the defined care priorities.

Intracranial hypertension can be the result of a primary central nervous system process, or it can result as a complication of a concurrent systemic illness. Independent of the specific etiology, it is an important and potentially disabling and deadly complication. While it often has important prognostic implications, its successful management can protect the brain from secondary injury and improve outcome.The management of intracranial hypertension is frequently approached in an overly simplistic manner with the sequential application of various intracranial pressure (ICP)–lowering strategies in a cookbook fashion. However, such a simplistic approach to intracranial hypertension can be perilous and potentially aggravate the problem. The management of intracranial hypertension is best tailored to each specific situation. This chapter will introduce the basic principles that should be considered when arriving at an individualized strategic management plan for a patient with brain swelling and/or intracranial hypertension. While some of these concepts and treatment strategies are discussed in Chap. 93 in the context of head injury, we will present a more global view on the management of brain swelling and intracranial hypertension

Intracranial Pressure

When it is monitored, the ICP tracing is a ballistic waveform much like the systemic arterial pressure (Fig. 65-1). However, it has a narrow “pulse pressure” and is expressed, by convention, as its mean. The normal mean ICP is generally between 5 and 10 mm Hg, and it will fluctuate at times to higher levels depending on many physiologic factors.

In normal adults, the skull is a rigid container and contains brain parenchyma, fluid (interstitial and cerebrospinal), and blood (arterial and venous). The expansion of any of these compartments or the addition of a space-occupying process can lead to ICP elevation. Table 65-1 lists some of the most common causes of brain swelling. The extent of the ICP elevation depends on many factors, most importantly the patient's intracranial compliance.

Intracranial Compliance

Figure 65-2 shows the intracranial pressure-volume curve that diagrammatically describes the dynamic of intracranial compliance. When a process that adds volume to the intracranial cavity evolves, cerebrospinal fluid (CSF) from the hemispheric convexities, basal cisterns, and ventricles (except in hydrocephalus) is passively displaced into the contiguous spinal subarachnoid space. As a result, the ICP only modestly increases during the early stages of the process (A). However, once CSF cannot be passively displaced any further, the ICP rises more sharply as the space-occupying process continues to evolve (B). This relationship between added volume to the intracranial cavity and its matched rise in ICP is referred to as the intracranial compliance. There is decreasing intracranial compliance as the intracranial hemodynamic relationship shifts to the right along the intracranial pressure-volume curve.

The concepts of the intracranial pressure-volume relationships and intracranial compliance can be applied when evaluating brain imaging studies to estimate the likelihood of intracranial hypertension in a patient who suffers a space-occupying intracranial process. For example, Fig. 65-3 shows an axial computed tomography (CT) scan of the brain with brain swelling from a middle cerebral artery territory infarction. As there are still compressible spaces around the swollen brain (e.g., basal cisterns, ipsilateral lateral ventricle, and ipsilateral cortical sulci) on that scan, it is reasonable to expect that the ICP is not yet significantly or persistently elevated. On the other hand, the CT in Fig. 65-4 shows a more extensively swollen brain than in Fig. 65-3, with obliteration of all surrounding CSF spaces. In the example shown in Fig. 65-4 the ICP is likely markedly elevated. While estimating the likelihood of intracranial hypertension by CT scan appearance is imperfect, it can provide some practically useful direction in management when the clinician is making decisions about whether to begin invasive ICP monitoring or considering other therapeutic maneuvers such as surgical decompression.

Cerebral Blood Flow and Cerebral Perfusion Pressure

Alterations in cerebral blood flow (CBF), either too much or too little, can disturb the brain homeostatic mechanisms and lead to cerebral injury. A relatively constant supply of blood to the brain is ensured by cerebral autoregulation, a physiologic phenomenon that allows maintenance of a constant CBF over a wide range of arterial blood pressures. Figure 65-5 graphically depicts this relationship for normotensive patients. In chronic hypertensives, the autoregulatory thresholds are shifted to the right. Relative blood pressure lowering within an autoregulatory range leads to cerebral vasodilation and an increase in cerebral blood volume. Conversely, relative blood pressure elevation within an autoregulatory range leads to cerebral vasoconstriction and a decrease in cerebral blood volume. Such alterations in cerebral blood volume can importantly affect the ICP. In fact, CBF is a critical determinant of ICP, and its reduction is a major final common pathway for cerebral injury from raised ICP. A practical understanding of these concepts is critical to developing a rational therapeutic plan for patients with intracranial hypertension.

The concept of cerebral perfusion pressure (CPP) is an important guide to management of the dynamic interplay between ICP and blood pressure. By definition, the CPP refers to the difference between the mean arterial pressure (MAP) and the cerebral venous pressure. Since the cerebral venous pressure is approximated by the ICP in most cases, the generally accepted CPP equation is:

This equation emphasizes the important point that the primary mechanism of cerebral injury from intracranial hypertension is cerebral ischemia from compromised cerebral perfusion. Notably, CBF is not a direct correlate of CPP, since autoregulation allows the CBF to remain constant in normotensive patients once the CPP exceeds approximately 40 mm Hg in previously normotensive patients.

The physiologic relationship between blood pressure, CBF, and CPP are less predictable in damaged brain regions where autoregulation is impaired. In brain regions without intact autoregulation, relative blood pressure lowering or elevation, even within a normal range, can lead to regional cerebral ischemia or hyperemia, respectively. Table 65-2 lists several important factors that influence CBF and ICP. Control or manipulation of these factors constitutes the basis for much of the medical management of raised ICP.

In most situations, it is best to maintain CPP in the 60 to 80 mm Hg range in euthermic patients, but the target CPP must be adjusted based on the priorities dictated by each particular clinical situation. For example, when brain hyperemia is one of the causes of brain swelling, maintaining CPP at a lower level can be an important therapeutic strategy. On the other hand, maintaining CPP on the generous side can provide a useful margin of “reserve” when there are plateaus in ICP that can suddenly and unpredictably challenge brain perfusion. Furthermore, higher CPP should be tolerated in circumstances in which blood pressure augmentation is used in selected patients as a treatment strategy for raised ICP based on its ability to cause autoregulatory reflex vasoconstriction, reduced cerebral blood volume, and ICP lowering.

Plateau Waves

One of the most feared and devastating complications of patients with intracranial hypertension and poor intracranial compliance are plateau waves (PW). PW are acute elevations in ICP, at times in the range of 50 to 100 mm Hg, in patients with reduced intracranial compliance. They can last from several minutes to more protracted durations in extreme cases. Resolution of a PW is often as abrupt as its occurrence. While there are many causes of PW, one important mechanism is generalized cerebral vasodilation from an autoregulatory response to a decrement in systemic blood pressure.1 Other causes include processes that can cause an increase in CBF and/or cerebral blood volume, many of which are listed in Table 65-2. Independent of their mechanism, PW can be disabling or deadly. Since compromised CPP can play a role in the occurrence of the most severe PW, relative blood pressure lowering should be avoided and/or rapidly treated. Similarly, during a PW, maneuvers that accentuate CPP such as blood pressure augmentation will abort the process in many circumstances. Even if blood pressure augmentation does not abort the process, it can prevent cerebral ischemia until other treatment modalities can successfully lower the ICP.

Cerebral Edema

Cerebral edema is defined as an increase in brain water content. The more classic categorization of cerebral edema by etiology differentiates that caused by cellular injury (cytotoxic edema) vs. breakdown of the blood-brain barrier (vasogenic edema).3 Hydrocephalic edema, ischemic edema (a combination of cytotoxic and vasogenic edema), osmotic edema, and hydrostatic edema have also been characterized as distinct entities by some authors based on their mechanisms and the location of the edema fluid.3–5Table 65-3 lists various categories of cerebral edema along with some of their distinguishing characteristics.

Vasogenic edema is comprised of a plasma-derived protein-rich exudate due to an alteration in the blood-brain barrier. It occurs in both gray and white matter, but it tends to predominate within the white matter. Vasogenic edema occurs in variable degrees with brain tumors, abscesses, traumatic brain injury, meningitis, infarction, and hemorrhage. Corticosteroids are useful in reducing vasogenic edema, and their effect is most profound when vasogenic is the main edema etiology, as with brain tumors, and to a lesser degree abscesses.6 Osmotic agents (see below) have no beneficial effect on the formation or extent of vasogenic edema.3

Cytotoxic edema is characterized as an intracellular swelling of neurons, glia, and endothelial cells with an accompanying reduction in the extracellular space. It occurs without a disruption in the blood-brain barrier and is likely due to cellular energy depletion. This results in the failure of the adenosine triphosphate–dependent sodium pump, with resultant accumulation of sodium and water within the cells.2,3 It occurs in both gray and white matter. Hypoperfusion injuries are most classically associated with cytotoxic edema. They have also been described with Reye's syndrome, diabetic ketoacidosis, and water intoxication. Corticosteroids have no benefit in limiting cytotoxic edema, and osmotic agents have only limited benefit in reducing brain water from cytotoxic edema due to concurrent disruption in autoregulation with most of the processes that cause this form of swelling.3

Ischemic edema begins as a cytotoxic edema, and is followed to variable degrees by vasogenic edema.4 When severe ischemia occurs and is followed by reperfusion, edema formation is biphasic: cells initially take up water and sodium from the extracellular space, but there is no net increase in brain water because there is no blood flow to serve as a source of water. Hours to days after the ischemic insult, the blood-brain barrier becomes disrupted, resulting in increased permeability and rapid accumulation of vasogenic edema. There is no therapy of demonstrated benefit for this process, and steroids have not been shown to be helpful, even though one aspect of this form of edema is vasogenic.7

Hydrocephalic edema is defined as an increase in brain fluid caused by the blockage of CSF flow pathways. It occurs in the periventricular white matter in association with hydrocephalus. In this situation the blood-brain barrier remains intact and the edema is in the extracellular space and of the same composition as CSF. Treatment consists of relieving the obstruction to CSF outflow or providing an accessory outflow pathway with a ventricular catheter or shunt.

In general, cerebral edema does not directly affect neural activity. Its main and most severe consequences are due to its mass effect and distortion of surrounding brain tissue. This can cause regional ischemia and result in the development of pressure gradients, leading to devastating brain tissue shifts.8–10

Brain Tissue Displacement and Herniations

It is important to differentiate mass effect and brain tissue displacement (BTD) from intracranial hypertension. As should be apparent from Fig. 65-2, there can be a substantial amount of mass effect without an important global elevation in ICP, and the mass effect alone can cause brain damage through its regional effect on brain perfusion and/or brain tissue displacement (e.g. herniation). BTD is a distortion of brain anatomy, and depression of consciousness is one common sign associated with such distortion. However, depressed consciousness can be a late accompaniment to BTD. Autonomic changes such as increasing blood pressure and exaggeration of the respiratory sinus arrhythmia are also changes that can occur early with evolving BTD, and recognition of these signs can serve as an early warning of this life-threatening phenomenon.11 The consequences of BTD depend on the extent and duration of the process, and its occurrence alone should not deter attempts to reverse the process.

A more thorough discussion of the clinical signs associated with the various herniation syndromes is covered in Chap. 67 but it is useful to note the relationship between horizontal brain displacement and depressed consciousness. With acute intracranial mass lesions, lateral displacement of posterior brain structures (diencephalon and brainstem) correlate with depression of consciousness. This can be practically quantified by measuring horizontal shift of the pineal gland on noncontrast CT scans. In the classic study describing this important relationship, horizontal shift of the pineal (from midline) by 0 to 3 mm correlated with alertness; 3 to 4 mm with drowsiness; 6 to 8 mm with stupor; and >8 mm with coma.12 This has clinical applicability in determining the cause of depressed consciousness in individual patients. When the extent of horizontal shift does not readily explain a patient's level of consciousness, other potential etiologic factors should be considered.

The falx cerebri is the dural structure that sagitally divides the left and right hemispheres. The anterior cerebral arteries course inferiorly and parallel to the falx cerebri, and they supply blood flow to the anterior, inferior, and medial frontal lobes. When there is BTD anteriorly, it can lead to herniation of brain tissue under the falx cerebri, subfalcine herniation, thereby compressing the anterior cerebral arteries and placing their territory of supply (inferomedial frontal lobes and caudate nucleus) in jeopardy for infarction (Fig. 65-6). Furthermore, subfalcine herniation can lead to obstruction of the outflow of the lateral ventricles (foramina of Monro), leading to obstructive hydrocephalus (usually contralateral to the mass lesion). Since subfalcine herniation is an anterior process, it may not lead to depressed consciousness unless it is very extreme.

The tentorium cerebelli is the dural structure that divides the supratentorial from the infratentorial space. The space between the lateral midbrain and the medial border of the tentorium cerebelli is called the tentorial incisura, and the posteromedial temporal lobe sits just above this space. When an intracranial mass or brain swelling is posterior it can cause BTD (mainly the inferomedial temporal lobe) through the tentorial incisura, called lateraltranstentorial herniation. This can lead to compression of the posterior cerebral arteries as they course around the midbrain, placing the brain tissue they supply (occipital lobes, medial temporal lobes, and thalami) in jeopardy for infarction (see Fig. 65-6).

Central transtentorial herniation (Fig. 65-7) is most common with global processes (e.g., global ischemia/infarction, meningitis, or fulminant hepatic failure) and classically occurs as the cerebral hemispheres and basal ganglia exert downward pressure, causing BTD through the tentorial incisura bilaterally. If persistent and extreme, it results in severe brainstem compression with resultant ischemia and hemorrhage, bilateral posterior cerebral artery compression with resultant ischemia/infarction in its territory of distribution, and CSF outflow obstruction with resultant hydrocephalus.

Posterior fossa lesions produce their effects through direct pressure on the medulla, or downward displacement of the cerebellar tonsils (cerebellar tonsillar herniation) and lower brainstem through the foramen magnum (see Fig. 65-7). Clinically, these patients develop early depressed consciousness, autonomic disturbances, altered respiratory patterns, and signs of brainstem dysfunction. This type and location of BTD causes its damage by severe brainstem and upper cervical spinal cord compression, as well as obstruction of CSF outflow.

As already mentioned, BTD from regional brain swelling can occur even without important ICP elevations, and it can have devastating consequences. For an example, an acute middle cranial fossa process such as an acute temporal hematoma can cause lateral transtentorial herniation without a profound rise in ICP. As a result, successful management of BTD requires strategic consideration of its significance in individual cases. This is particularly important since the treatment for elevated ICP is not necessarily the same as it is for BTD. In fact, in certain cases ICP-directed medical treatments can accentuate BTD.

When a patient at high risk for brain swelling is encountered, a proactive approach to management should begin. This includes developing a monitoring strategy for early detection of the problem and planning the best management approach if it does occur. In at-risk patients without significant brain swelling, invasive ICP monitoring may not yet be appropriate. However, early insertion of an ICP monitor may be necessary if the risk of the problem is very high (e.g., fulminant hepatic failure), serial examination and/or imaging cannot be performed properly or readily (e.g., the intubated and/or heavily sedated patient), and/or the determination of the ICP will translate into important therapeutic changes.

It should be clear that the most rational and sophisticated management of intracranial hypertension should be based on continuous ICP and CPP monitoring. While active management of ICP guided by ICP monitoring has not shown a consistent benefit in outcome, it is difficult to study this well in humans for a variety of methodologic reasons. Most experienced physicians appreciate that ICP monitoring has value in properly selected patients. In particular, processes associated with life-threatening elevation of ICP, a rapidly expanding intracranial mass, or a disease in which ICP elevation is likely to be progressive and demand proactive preservation of CPP represent some generally accepted indications for monitor insertion.

ICP can be monitored from several intracranial sites (Fig. 65-8). The most commonly employed ICP monitoring devices are ventricular catheters and parenchymal monitors. Both can be readily placed at the bedside with a twist drill by neurosurgeons or experienced neurocritical care physicians. Ventricular catheters have been considered the gold standard for ICP monitoring, based on their reliability and ability to drain CSF, as an alternative for ICP control. It is the most invasive of alternatives, however, and it is associated with higher risks of infection, intracranial hemorrhage, and in our experience, seizures.

Ventricular catheters are ideally placed into a lateral ventricle. In general, right-sided (nondominant hemispheric) placement is usually preferred, since a bleeding complication from the insertion would be less likely to cause language disturbance from this location. Ventricular catheters are then connected to one of the many commercially available closed external drainage systems comprised of a drainage chamber, drainage bag, and an attached transducer. The height of the drainage chamber can be set to allow automatic drainage when the ICP rises above a designated level based on manometric concepts. A ruled measure on the drainage system facilitates accurate drainage chamber height adjustments and a stopcock system between the drainage catheter, drainage chamber, and transducer allows the physicians and nurses to choose between automatic CSF drainage and continuous ICP transduction.

One behavior we often encounter is the simultaneous transduction and recording of the ICP while the CSF chamber is set and open to automatically drain. In this situation, the ICP measure obtained is generally meaningless. For this reason, even when continuous CSF drainage to a determined level is the selected care strategy, the ICP should be transduced with the ventricular drain closed to drainage at least every 30 to 60 minutes for a 5- to 10-minute period with a recording of the ICP and CPP at the end of that time and determination if they are within the desired parameters. This is particularly important since CSF drainage to a set level does not always achieve maintenance of the ICP in the desired range. Some of the factors that limit ICP control by this method are imperfect ventricular catheter placement, collapse of the ventricle in which the catheter is placed, bubbles within the catheter or fluid column from catheter to transducer, and catheter obstruction.

Several avoidable causes of complications from ventricular drains should be mentioned. Ventricular drains should never be connected to a pressure bag. This can be a deadly error. The only connections to the ventricular catheters should be the drainage system. The drainage system should not be detached or penetrated with a needle except by individuals experienced with the management of ventricular drains. Ventricular drains should be actively closed (via the intervening stopcock apparatus) whenever a change is being made in patient position (particularly involving raising the head of the bed) or during excessive coughing or active endotracheal suctioning. Failure to do this can lead to excessive CSF drainage with dangerous consequences. In addition, ventricular catheters are structurally fragile and they can be readily broken or displaced from their placement within the cranium. All personnel moving patients with ventricular drains (or all other types of ICP monitors or intracranial drains) should be aware of their presence and reminded to take special care in the process. When transporting patients, assigning one individual to have responsibility for the head and maintenance of the monitor integrity is advisable to avoid inadvertent mishaps.

An important risk of infection is associated with their use, and its occurrence in properly handled drains is most often related to its duration of use. Catheter exchange does not have a demonstrated benefit in reducing the infection risk when prolonged external drainage is required, and the practice expands the risk of iatrogenic complications from multiple insertions. Longer subcutaneous catheter tunneling and giving prophylactic antibiotics that target common skin flora likely decrease the risk of infection.13,14 While newer antibiotic-coated catheters promise a lower risk of infection, their use is not presently an established care standard in most medical settings. Due to the risk of infection with ventricular catheters, each day a patient has one in place should be used to maximize benefit from its placement and/or begin the process necessary to determine the best timing for its safe removal. For a patient on subcutaneous heparin or heparinoids for deep vein thrombosis prophylaxis, it is our practice to hold two doses before removing the ventricular catheter to hypothetically decrease the risk of development of hematoma during removal.

Patients with ventricular drains should have their CSF analyzed at least every other day (we do so daily) for white blood count, protein, glucose, Gram stain, and bacterial cultures. A reduction in CSF glucose to <50% of the simultaneous serum glucose is ominous for infection, particularly when accompanied by a rising CSF white blood cell count, fever, nuchal rigidity, and worsened encephalopathy. However, nuchal rigidity, fever, and worsened encephalopathy can be later findings, and their absence should not delay a real suspicion for early CSF infection. Conversely, a mild elevation in CSF white blood count alone has limited specificity since many processes that require a ventricular drain (e.g., hemorrhage) cause a delayed inflammatory response. We have observed that relative lowering of serum sodium occurs in a respectable number of patients as an early sign of evolving CSF infection in those with ventricular drains. In general, compulsive concern about CSF infection should drive tight monitoring for its occurrence and early empiric treatment when suspected.

Parenchymal monitors offer some advantages over other systems in that they allow rapid and safe placement with accurate measurements and less risk of infection. The presently available parenchymal monitors employ either fiberoptic or strain gauge technology. Their tip can be placed anywhere within the skull, but their most common site is several millimeters into the brain parenchyma through a twist drill hole placed over the convexity. Since the maximal ICP is in the region of an intracranial mass, they are best placed ipsilateral to a mass lesion unless it is deemed unsafe based on the suspected etiology of the process.

Parenchymal monitors are often placed through and secured by a bolt screwed into the cranium after the placement of a small hole with a twist drill; avoidance of kinking is important for fiberoptic catheters since they can break if bent. The monitors based on strain gauge technology can be bent and allow a wider flexibility of placement methods and configurations relative to the fiberoptic systems. However, both technologies are widely used and the selection of monitor type is often based on institutional history and the personal experience and preferences of those who insert them. While ventricular catheters are fluid coupled systems that can be zeroed by nurses at the bedside, parenchymal monitors are zeroed prior to insertion and can only be rezeroed through removing and reinserting them. The advantages of parenchymal monitors over ventricular catheters are to some degree counterbalanced by drift in the ICP readings over time and the inability to drain CSF for ICP control.15 While these various advantages and disadvantages are important considerations when selecting the ICP monitor type, the anatomy and physiology of the process requiring a monitor often dictate the best monitoring method.

As illustrated in Fig. 65-1, ICP monitors provide a ballistic waveform temporally associated with the systemic blood pressure. Certain characteristics about the ICP waveform can communicate important information about an evolving cerebral process, even when the ICP is not critically elevated. The ICP waveform is comprised of various smaller subwaves. The first two waves observed during systole are referred to as P1 and P2. Normally, each sequential subwave of the ICP waveform is smaller than the one before it, so P1 is usually greater than P2. However, as intracranial compliance decreases, P2 usually becomes greater than P1, even before the ICP is increased to abnormal levels. In addition, fluctuations in the baseline of the ICP waveform with breathing usually reflects a decrease in intracranial compliance, and it can occur and accentuate even before an elevation in ICP. These facts can be helpful in selected situations. For example, when a patient has fulminant hepatic failure, observation of these subtle changes in the ICP waveform can signal the evolution of brain swelling before it is clinically apparent or the ICP has increased.

As mentioned earlier in this chapter, ICP elevation can be imperfectly predicted through analysis of brain imaging studies, usually CT. When the ICP from any monitor type is either much lower or higher than expected clinically and based on imaging analysis, steps should be taken to determine the accuracy of the monitor. At times this requires rezeroing the monitor or replacement, at times even in a new position or location. Blind trust of monitoring information that is significantly contrary to clinical expectations can lead to mismanagement and avoidable complications.

While Chap. 93 presents much about the available conventional management strategies for intracranial hypertension when it complicates traumatic brain injury, they are similar for nontraumatic etiologies. The following sections will supplement the information provided in that chapter.

Controlling Aggravating Factors

Patient positioning is an important aspect of the care of patients with intracranial hypertension. Lateral head rotation, rotating neck movements, and restrictive devices around the neck (e.g., endotracheal tube neck straps) can compromise jugular venous drainage. The head should be kept in the neutral forward position. Sometimes towel rolls placed on both sides of the head can be useful in maintaining the neutral position. Head elevation lowers ICP by enhancing CSF drainage and maximizing cerebral venous output. CSF and cerebral venous blood are maximally displaced with severe intracranial hypertension, challenging the benefits of head elevation in these patients. The logical approach is to individualize the ideal head and neck position based on clinical observation of the patient and ICP monitoring, when available. When ICP monitoring is not available, head elevations of 30° to 40° allows the most consistent reduction in ICP.16

Fevers should be treated aggressively because body temperature elevations increase ICP by increasing cerebral metabolism, cerebral blood flow, and cerebral edema.17,18 Acetaminophen should be administered (assuming normal hepatic function) and a diligent search made for treatable infection. There should be a low threshold for using empiric antibiotics for the most likely source, even though centrally mediated fevers are often considered in the differential diagnosis in many of these patients. In those resistant to acetaminophen, nonsteroidal anti-inflammatory drugs (e.g., ibuprofen) should be considered when bleeding is not a primary concern. Cooling blankets can be useful, but control of shivering is necessary since it can aggravate intracranial hypertension. Newer intravascular cooling devices provide an exciting alternative to more rapidly and consistently achieve fever control in this brittle patient group.19 Independent of the cooling method, if shivering complicates its use, it can be controlled with various medications such as meperidine, parenteral sedatives, and neuromuscular paralysis. The key point about fever control in patients with brain swelling and/or intracranial hypertension is that it is essential and requires close follow-up to assure that it is brought under control in a timely fashion.

Coughing and straining against the ventilator increases intrathoracic pressure and can reduce the venous outflow from the intracranial cavity. This can lead to elevation of the ICP. As a result, strategies to decrease these ventilator- or airway-provoked problems can be essential to the management of patients with brittle intracranial hypertension. Some of these strategies include repositioning of the endotracheal tube, inhaled nebulizations that can serve to decrease coughing, ventilator adjustments adapted to the patient's respiratory behavior, or the administration of sedation or nondepolarizing neuromuscular blockers.

Positive end-expiratory pressure (PEEP) will increase ICP only when mean airway pressures are increased, causing transmission to the mediastinum. When pulmonary compliance is reduced as occurs in acute respiratory distress syndrome or pneumonia, the effect of PEEP on ICP is attenuated. It seems fortuitous that the conditions for which higher PEEP levels are often used diminish its deleterious impact on ICP. Probably more important than the effect of PEEP on ICP is its potential effect to decrease cardiac output and blood pressure, with their deleterious impact on cerebral perfusion. In general, the presence of an ICP monitor shifts the concerns about the impact of PEEP from the theoretical to the empiric.

Frequent zealous monitoring of systemic blood pressure is a cornerstone of ICP management in patients with intracranial hypertension. Hypotension will directly cause cerebral vasodilation and increase ICP. A treatment plan should be orchestrated with pressor agents and isotonic fluids at the bedside to allow rapid intervention at the first observation of significant relative blood pressure lowering. However, when cerebral autoregulatory mechanisms are impaired, hypertension may lead to pathologic increases in regional CBF, worsened brain swelling, and aggravation of intracranial hypertension.

Pain and arousal can cause elevated ICP by increasing CBF. As a result, patients with life-threatening intracranial hypertension should generally receive sedatives and analgesics, even if their use risks clouding of the neurological exam. The present availability of short-acting sedatives and analgesics and the ready availability of neuroimaging have simplified the historical reluctance to utilize these classes of agents in the neurologically ill.

Seizures increase ICP by increasing cerebral metabolism and CBF by Valsalva. If the patient is deemed to be at risk, anticonvulsants should be administered promptly. Given the deleterious consequences of seizures and the severity of illness of these patients, drugs that can be parenterally administered with known effectiveness as monotherapy should be used (phenytoin, phenobarbital, or valproic acid). Since hypotension can accompany the administration of phenytoin and phenobarbital, close monitoring of the blood pressure is critical, with preparedness to detect any relative blood pressure lowering and promptly intervene.

Mechanical Ventilation and Hyperventilation

Hyperventilation induces a rapid and effective reduction in ICP through vasoconstriction induced by hypocapnia-associated CSF alkalosis.20 The duration of the ICP reduction is variable, but in general, the ICP returns to baseline within hours after commencing hyperventilation, due to normalization of the CSF alkalosis through compensatory adjustments in the bicarbonate buffering systems in the brain and vascular smooth muscle.

When hyperventilation is required for urgent management, it can be accomplished with an ambu mask or mechanical ventilation. Providing a 10- to 12-mL/kg tidal volume at a rate of 14 to 20 breaths/min usually achieves substantial reduction in the partial pressure of carbon dioxide (PCO2). The ideal PCO2 is variable depending on the clinical situation and the individual patient's response. One of the most controversial questions is whether excessive hyperventilation can cause cerebral ischemia through extreme cerebral vasoconstriction—a phenomenon suggested by several studies in traumatic brain-injured patients.21 The most recent work on the subject in patients with severe traumatic brain injury utilizing positron emission tomography demonstrated that hyperventilation can be safely performed (without consequent cerebral ischemia) to a PCO2 of 30 mm Hg, and perhaps to <25 mm Hg in selected patients.22 Furthermore, it has been shown that hyperoxia can improve oxygen delivery to the brain during hyperventilation.23

The potential benefits of hyperventilation must be balanced against some of its potential deleterious consequences, including but not limited to diminished cardiac filling pressures with resultant hypotension, decreased myocardial oxygen supply with an increase in myocardial demand, elevation in mean airway pressure leading to accentuation of intracranial hypertension, electrolyte disturbances (e.g., alkalosis, hypokalemia, and hyperchloremia), and cardiac arrhythmias.24

Because hyperventilation is the most rapid way to reduce ICP, it is best kept as a back-up strategy for emergent ICP elevations. A modest goal of 30 mm Hg seems safest with the present state of our knowledge, and preoxygenation may add a layer of protection that is potentially beneficial for ischemia deriving from hyperventilation. Once other ICP-lowering strategies are instituted and controlling ICP and CPP adequately, hyperventilation should be lifted. Gradual withdrawal of hyperventilation is necessary to avoid rebound elevations in ICP as the PCO2 is normalized. We recommend not exceeding PCO2 increases of 2 to 3 mm Hg per hour. When hyperventilation is used for a prolonged period, jugular venous oxygen saturation can be used to monitor for the feared complication of cerebral ischemia due to this maneuver. Inadvertent fluctuations in the PCO2 due to variable ventilation is an important and avoidable cause of ICP plateaus. This is commonly a problem that occurs during patient transport. We recommend using transport ventilators for patients with brittle intracranial hypertension to minimize variations in PCO2 during transport.

Tris(hydroxymethyl)aminomethane (THAM) is a buffer than can be used to correct acidotic states, and it can be used at times to assist in the management of patients with intracranial hypertension. The advantage of THAM is that it alkalinizes without changing plasma sodium or PCO2. THAM may have a role in limiting rebound ICP elevation during the withdrawal of hyperventilation, or prolonging the benefit of hyperventilation in some patients.25 It is administered intravenously at a dose of 1 mL/kg per hour. Some of the complications associated with its use include local skin irritation and necrosis, hypoglycemia, and respiratory depression.

Cerebrospinal Fluid Drainage

CSF drainage with a ventricular catheter is the primary treatment for hydrocephalus, and it can be a useful treatment strategy for intraventricular hemorrhage, through both CSF drainage and providing access for administration of intraventricular thrombolytic agents. The management of CSF drainage and ventricular catheters is discussed in the section on ICP monitoring in this chapter. While CSF drainage through lumbar puncture should generally be avoided in patients with intracranial hypertension, it is a mainstay of treatment for patients with pseudotumor cerebri (benign intracranial hypertension).

Mannitol

Osmotherapy is directed at increasing plasma osmolality, establishing an osmotic gradient across the relatively impermeable blood-brain barrier. This favors a net loss of brain water, thereby increasing brain compliance and decreasing ICP. In addition, mannitol, the most widely used osmotic agent, can improve cerebral perfusion through transient hypervolemia and hemodilution, leading to autoregulatory cerebral vasoconstriction, decreased cerebral blood volume, and lower ICP.26–28 It also can increase CSF absorption.29 However, most mechanisms of brain injury lead to disruption of the blood-brain barrier and a loss of normal autoregulation. As a result, osmotic agents are less effective in injured brain regions. This is an important consideration when strategizing management in individual patients with unilateral mass lesions and BTD.

Mannitol is typically used in a 20% solution. Its ICP-lowering effect is dose-dependent, and it appears to be maximal with a 1 g/kg dose infused over 30 minutes.30,31. The duration of benefit from this high-dose infusion is limited to several hours, and terminal dose-interval ICP elevations beyond the pretreatment baseline are common. Smaller doses of 0.25 to 0.5 g/kg over longer infusion periods (30 to 60 minutes) can be associated with less profound and more lasting ICP benefits within the usual 4-hour dosing interval when multidosing regimens are used. The reduction in ICP after mannitol use should be apparent within 15 minutes, and failure of a response to mannitol is considered ominous, but not hopeless, in our experience.

The goal of osmotherapy should be plasma hyperosmolality with maintenance of adequate intravascular volume. In general, the target osmolality should be in the 300 to 310 mOsm/L range with adequate maintenance intravenous fluid administration. Normal saline should be used for maintenance and replacement fluids. Hypotonic fluids are contraindicated and should be avoided because water accumulation by hyperosomolar brain can aggravate brain swelling and intracranial hypertension. The development of moderate hypernatremia should be expected and tolerated, because administration of hypotonic fluids to correct the hypernatremia is counterproductive and potentially dangerous.

Single doses of mannitol are often effective in reducing ICP and improving intracranial compliance temporarily. When single doses are used, monitoring of the plasma osmolality is unimportant. The benefit of multidosing regimens is more controversial. This is due to some experimental evidence that it can accentuate regional brain water when the blood-brain barrier is disrupted.32–34 In addition, its disproportionate impact on uninjured brain (relative to injured brain) can potentially accentuate BTD through enhancement of regional intracranial pressure differentials.35 Nonetheless, mannitol is an effective agent for rapidly reducing ICP in an emergency setting, and it can buy critical time to temporarily stabilize the patient while other treatment strategies are being arranged and commenced.

Rapid diuresis before administration of adequate amounts of replacement fluids can occur in some patients after receiving mannitol. This can lead to systemic hypotension, emphasizing the need for preparedness when using this treatment modality. We always have pressor agents and replacement fluids readily available at the bedside in patients with increased ICP. Other important complications include electrolyte disturbances (hypernatremia and hypokalemia), prerenal azotemia, and congestive heart failure. Sometimes, pretreatment with a dose of furosemide can circumvent some of the complications of the transient expansion of intravascular volume that occurs immediately after administration of mannitol.

Hypertonic Saline

Hypertonic saline has been used with renewed enthusiasm in patients with brain swelling and intracranial hypertension.36–39 The osmolality of 3% saline (1026 mOsm/L) is similar to that of 20% mannitol (1375 mOsm/L). The osmolality is much higher with other solutions, such as 7.5% saline (2565 mOsm/L) and 23.4% saline (8008 mOsm/L). Variable formulations of hypertonic saline have been used with different-size boluses (up to 75 mL) with occasional benefit, most recently reported in some patients with traumatic brain injury and cerebral infarction refractory to other modalities. Its use can be associated with the expected complications of hypernatremia, hypokalemia, hyperchloremia, and congestive heart failure. It is an underrecognized part of the ICP-lowering armamentarium. Its use can be a particularly well-suited treatment in patients with volume depletion or renal failure.

Loop Diuretics

Loop diuretics such as furosemide can reduce ICP and have been reported in selected series to reduce ICP alone or when used in conjunction with osmotic agents.40 Diuretics exert their effect through an osmotic gradient caused by a mild diuresis, reduction in CSF formation, and reduction in brain water. They provide another alternative when mild reductions in ICP are desired. However, like osmotic agents, they can produce volume and electrolyte depletion, requiring proper treatment and surveillance for these complications.

Corticosteroids

Corticosteroids have been shown to reduce the vasogenic edema associated with brain tumors and abscesses, but their benefit with other processes that can cause intracranial hypertension is less clear.41 There is generally no benefit consistently demonstrated with intracerebral hemorrhage.42–45 While it may have a theoretical advantage in certain stages of infarction-related edema, it has not been of demonstrated benefit and has been associated with worse outcomes.46–49 Dexamethasone is the most commonly used corticosteroid for vasogenic cerebral edema, and there is a wide range of dosing alternatives. We typically use 16 to 24 mg/day of dexamethasone in two to four divided doses in extreme situations. It can be given parenterally or enterally. Higher doses can be safely used for brief periods of time with less clear benefit over more conventional dosing. Potential side effects include gastrointestinal bleeding, hyperglycemia, disturbance in nitrogen metabolism, wound breakdown and poor healing, sleep disturbances, and behavioral disturbances. The hyperglycemia is of particular concern and requires aggressive intervention due to its increasingly recognized negative impact on outcome with most of the processes that can cause intracranial hypertension.

Hypothermia

There has been renewed interest in moderate hypothermia (32° to 34°C) as an adjunctive therapy for patients with intracranial hypertension. It can lower ICP and improve CPP in some patients, and it theoretically limits hypoperfusion brain injury due to its neuroprotective properties. However, it has not yet been demonstrated to achieve outcome benefit in scientifically valid clinical trials. In spite of the lack of scientific evidence for its benefit to date, those experienced in the management of patients with the problem have used hypothermia with success in selected patients; it is considered a useful “time-buying” trick in refractory cases.

The most classic method for hypothermia induction is with cooling blankets placed above and below the patient with the variable addition of ice-water lavage. However, newer intravascular cooling devices are increasingly available but not yet widely applied.19 When hypothermia is used, shivering can be a complicating factor when the body temperature gets below 36°C. Since shivering can increase ICP, its occurrence should be anticipated and sedatives and/or narcotics should be rapidly administered and titrated to effect to limit its severity. Rebound intracranial hypertension is an important concern during the rewarming process.50 When induced hypothermia is terminated, patients should be allowed to passively rewarm with close attention to slow down its pace if the patient develops a rebound increase in ICP. The most common complications of induced hypothermia include pneumonia, cardiac arrhythmia, and coagulopathy.51

Barbiturate Coma

The use of barbiturate coma is an advanced treatment measure aimed at preventing patient demise from uncontrollable intracranial hypertension. While it has been used with variable degrees of success for the treatment of elevated ICP from a variety of causes, there is little evidence that it improves outcome.52–56 In addition, it is a costly undertaking from an economic and physiologic perspective.

Some of its ICP-lowering benefit may be derived from depression of cerebral metabolism and reduction of CBF in normal areas of brain, with shunting of blood to ischemic areas. In addition, it may limit oxidative damage to lipid membranes and scavenge free radicals, reduce formation of vasogenic edema, attenuate fatty acid release, reduce intracellular calcium, and of course limit external stimuli from causing patient arousal, with the concomitant increases in CBF and cerebral blood volume.48,54

The induction of barbiturate coma is a weighty undertaking and demands experience to assure its safe and proper use. The agents most commonly used are thiopental and pentobarbital. Systemic hypotension, complications of prolonged immobility and mechanical ventilation, and immune suppression are the most common deleterious effects of barbiturate therapy. In addition, the need for more frequent transport and invasive line complications are other risks that should not be underestimated in their potential impact.

Pentobarbital coma requires a loading dose of 10 to 30 mg/kg. In our experience, it is best to administer the pentobarbital in small boluses of 100 to 200 mg every 10 to 20 minutes as tolerated from a blood pressure standpoint. This should be done under electroencephalographic (EEG) monitoring. While each bolus will briefly achieve either a burst-suppression pattern or a flat EEG, usually a more full loading dose is necessary to achieve a sustained effect. A continuous drip of 1 to 3 mg/kg per hour is usually necessary to maintain the desired depth of anesthesia. Optimizing depth of barbiturate therapy should be guided by EEG in these cases, usually titrating a continuous drip to achieve a burst suppression pattern in the 3 to 6 bursts/min frequency. The EEG can be monitored continuously or hourly, and ICU nurses can be taught to interpret burst-suppression frequency as part of the bedside monitoring.

In patients under prolonged barbiturate anesthesia, various strategies must be used to compensate for the loss of ability to do a neurologic examination, and can include continued evoked potential monitoring, serial imaging studies, and numerous other devices that are increasingly becoming available that can track a number of combinations of intracranial factors. The only neurologic exam that is useful in these patients is pupillary dilation, since the barbiturates usually cause bilaterally small pupils. When they become large it can be an ominous sign. However, an important phenomenon has been described in patients under pentobarbital coma: an accentuated ciliospinal reflex that manifests as pupils that are large (>6 mm) and seemingly unreactive to light, usually after a nursing maneuver such as patient turning. It can be misinterpreted as a catastrophic clinical change leading to unnecessary scans. Usually, when it occurs the pupils will react with a sustained intense light stimulus, and it spontaneously abates within minutes after it begins.57

As already mentioned, intensive hemodynamic monitoring is required with barbiturate coma. Since volume depletion increases the risk of hypotension from barbiturates, special attention should focus on maintaining intravascular volume with the guidance of invasive monitoring in most cases. The risk of infection and the concurrent disruption in the usual response to infection—fever—requires systematic surveillance for infection with regular cultures (at least every other day) of the endotracheal secretions, urine, and blood drawn through invasive lines.

Hypothermia is common with barbiturates, and in fact we often use barbiturates to facilitate the induction of hypothermia as a treatment modality. While permissive moderate hypothermia is recommended when using barbiturates, extreme hypothermia (<32°C) is associated with numerous complications and should be avoided.

The long half-life of pentobarbital (approximately 24 hours) allows a slow recovery, even when abruptly stopped. Shivering is common during the recovery period from barbiturate anesthesia and may require treatment with narcotics or short-acting sedatives. We prefer to use propofol for this complication. In addition, chaotic EEG patterns are common during this period and often are misinterpreted as status epilepticus.

Surgical Procedures

Surgical procedures can be a very important part of the management of patients with intracranial mass lesions, with or without intracranial hypertension. The two primary types of procedures are removal of the primary problem (e.g., hematoma evacuation, neoplasm excision, etc), or creating more room to accommodate an intracranial mass and deter the deleterious effects of BTD by removing skull bone and opening the overlying dura ipsilateral to the mass through hemicraniectomy and durotomy. The application of these surgical alternatives and their timing is heuristic, based on numerous clinical factors, some of which will be discussed in the following section.

While we cannot provide an exhaustive delineation of management recommendations for all causes of intracranial hypertension, in this section we address some disease-specific recommendations for some of the causes most commonly encountered in a medical critical care setting, and not covered in other chapters of this textbook.

Hypertensive Encephalopathy and Eclampsia

Hypertensive encephalopathy occurs when the blood pressure is elevated beyond autoregulatory thresholds. This leads to an increase in extracellular water, predominantly by hydrostatic mechanisms. In addition, there can be variable degrees of parenchymal hemorrhage, most often localized in the end-arterial border zones along the frontal and posterior parietal convexities (see Fig. 65-9). This is an important reversible cause of brain swelling in which the extent of brain swelling does not necessarily correlate with neuronal injury. Blood pressure lowering is obviously the focus of treatment of the brain swelling with this disease. However, blood pressure lowering does not achieve immediate resolution of the cerebral edema. So, if the patient already has important intracranial hypertension and diminished intracranial compliance (an estimate guided by clinical and imaging parameters), ICP monitoring is necessary to guide the pace and degree of blood pressure lowering so as to not compromise cerebral perfusion. Any type of ICP monitor can be used, but most often global cerebral edema is associated with a loss of CSF spaces and compression of the ventricles. When that occurs, the ventricular catheters are more difficult to place, their readings are less consistently accurate, and they do not allow enough CSF removal to justify their higher risk. We find parenchymal monitors to be a more rational choice in these cases.

While the mechanism of brain swelling from eclampsia is similar but not exactly the same as hypertensive encephalopathy, the management of intracranial hypertension complicating eclampsia is generally the same. However, the management of eclampsia-associated intracranial hypertension has the added priority of urgent fetal delivery. The presence and suspected severity of intracranial hypertension should be considered when determining the method of delivery and anesthesia. Cesarean section is the preferred mode of delivery in almost all cases. Spinal anesthesia should be avoided due to the risk of precipitating central herniation with CSF drainage. General anesthesia should include close attention to the blood pressure to avoid degrees of lowering that could compromise cerebral perfusion. In general, successful management of intracranial hypertension is best guided with an ICP monitor, and easy and predominantly uncomplicated bedside placement of parenchymal monitors provides little justification for not using them.

Both hypertensive encephalopathy and eclampsia can be associated with ominous clinical presentations and imaging studies. Neither midposition unreactive pupils with extensor posturing nor CT changes suggestive of bilateral end-arterial border zone infarctions with hemorrhage should deter aggressive and optimistic management of such patients. In our experience, both scenarios can potentially lead to good outcomes when treated promptly and aggressively.

Fulminant Hepatic Failure

Cerebral edema complicates fulminant hepatic failure (FHF) in up to 50% of cases.58 Similarly to hypertensive encephalopathy and eclampsia, the brain swelling from FHF is global and usually symmetric (Fig. 65-10). Its etiology is thought to be predominantly glial swelling from intracellular glutamine accumulation, but a variety of other mechanisms have been shown to play a variable role in its development.59–62

The development of brain swelling from FHF should be anticipated in all patients. Early detection of cerebral edema is imperative in order to pace its development and proactively manage its complications. However, bedside clinical evaluation does not allow its accurate detection due to the worsening hepatic encephalopathy concurrently developing in FHF patients. For example, stage IV encephalopathy patients frequently have diffuse hyperreflexia and increased motor tone with decerebrate posturing in the absence of any cerebral edema. It is not possible to differentiate uncomplicated hepatic encephalopathy from FHF-associated cerebral edema in patients with higher-stage encephalopathy. However, significant cerebral edema most often occurs in patients with higher-stage encephalopathy (III or IV). CT scans should be done in all patients with FHF, particularly those with stage III or IV encephalopathy, to rule out intracerebral hemorrhage and to estimate the presence and severity of cerebral edema. In our experience, patients in the early stages of cerebral edema have their CT scans misread as normal by inexperienced physicians. In addition, the presence of intracranial hypertension has been reported in FHF patients with normal CT scans.63,64

Once a patient drifts into the higher stages of encephalopathy with FHF, significant cerebral edema and cerebral hypoperfusion can be occurring and be undetectable in the absence of ICP monitoring. So an ICP monitor should be placed prior to stage IV encephalopathy in FHF patients. Often the frequent and severe coagulopathy observed in these patients justifiably demands thoughtful reflection on the risks and benefits of the intervention. However, in our experience, even with coagulopathy, bedside parenchymal monitors can be inserted without significant complications. We prefer to use a parenchymal monitor in the nondominant standard location. We insert it with the administration of 2 units of fresh frozen plasma (FFP) before, 1 unit of FFP during, and 2 units of FFP after the procedure. This should be carried out in collaboration with other caregivers, so invasive line insertions can be coordinated while the FFP infusions are being administered for ICP monitor insertion. Small pediatric patients require a different strategy to minimize bleeding during invasive procedures. While some neuroclinicians prefer to insert ventriculostomies in patients with potential intracranial hypertension (to allow CSF drainage as one therapeutic maneuver to treat plateaus in ICP), this modality carries a higher risk of hemorrhagic complications and the ventricles collapse (not allowing CSF drainage) in patients with consequential cerebral edema from FHF.

Once an ICP monitor is in place, management should focus on maintenance of adequate cerebral perfusion pressure. Given the likelihood of impaired autoregulation in FHF patients, extremes of blood pressure should be avoided. The patient's blood pressure should be maintained in the mid-normal range for the previously normotensive patient. In addition, normal intravascular volume should be maintained, and a state of readiness for acute response to any systemic hypotension is critical.

In patients with an elevation in ICP and poor intracranial compliance, ICP lowering may be necessary, particularly those with tenuous CPP (<60 mm Hg). Mannitol can be useful in some patients with FHF.65,66 It should mainly be considered in nonoliguric/anuric patients without significant hypernatremia. We generally use 0.25- to 0.5-g/kg aliquots of mannitol infused over a 20 to 30 minute period. The maintenance intravenous fluid should be isotonic with serum; normal saline is a good choice, with added 5% dextrose and/or potassium, depending on the specific clinical situation.

When using osmotic agents with FHF, the patient is already in quite a hyperosmolar state, so we do not titrate the mannitol to a specific serum osmolality. Rather, it is used intermittently to achieve the desired affect of improved intracranial compliance and ICP lowering. It may lose its effectiveness after several dosings, and so other therapeutic strategies should be considered concurrently. As is typical with the use of an osmotic diuretic, volume depletion can be an important complication with resultant hypotension. Maintaining euvolemia is ideal and requires thoughtful attention to fluid balance and intermittent replacement. Administration of hypotonic fluids such as 50% dextrose without electrolytes (commonly used in FHF in some settings) should be avoided.

Most patients with FHF will develop spontaneous hyperventilation, and this can be related to an increase in circulating free fatty acids and ammonia. While the institution of hyperventilation can be beneficial in lowering ICP acutely, it has no prolonged benefit in FHF as is true with most etiologies of cerebral edema and intracranial hypertension.67

Induced hypothermia can be a favorable strategy for both neuroprotection and deterring the development of cerebral edema, but further research is required to clarify its preventive role in this population. A recent study supported the promising role of moderate hypothermia for lowering ICP in patients with FHF.68 Seven consecutive FHF patients with intracranial hypertension refractory to osmotherapy and ultrafiltration were studied. Moderate hypothermia to 32° to 33°C was achieved with cooling blankets. The mean ICP before and after cooling was 45 mm Hg and 16 mm Hg, respectively. The mean CPP before and after cooling was 45 mm Hg and 70 mm Hg, respectively. The mean CBF before and after cooling was 103 mL/100 g per minute and 44 mL/100 g per minute, respectively.

We induce hypothermia in patients with refractory intracranial hypertension prior to significant cerebral hypoperfusion events. We strive for a body temperature of 32° to 34°C with the use of cooling blankets above and below the patient. Proactive strategies to allow early detection and treatment of fever should be an important part of the care plan for these patients, independent of whether hypothermia is used.

Like most other therapeutic strategies for ICP lowering in FHF, there have been few reliable clinical studies to guide the indications for and use of barbiturate therapy. It has been shown to have an impact on controlling ICP in Reye's syndrome.69 It probably would be best considered in patients with either refractory intracranial hypertension and/or those with oliguria or anuria. One study of 13 FHF patients with acute renal failure and refractory intracranial hypertension administered thiopental slowly to a maximum of 500 mg to achieve an ICP <20 mm Hg, CPP >50 mm Hg, or until hypotension developed.70 In each case, the ICP was reduced with the administration of anywhere between 185 and 500 mg (median = 250 mg) of thiopental over a 15-minute infusion period. In eight patients, a constant infusion was required (50 to 250 mg/h) to maintain adequate ICP and CPP. Given the small number of patients and unclearly defined end points, it is difficult to assess the true benefit of the ICP-lowering accomplishments of this strategy. However, unique to FHF, impaired barbiturate metabolism and clearance often precludes the need for a maintenance infusion after the desired effect is accomplished using a loading dose. See the section on treatment recommendations for the management of barbiturate coma elsewhere in this chapter.

Global Cerebral Hypoperfusion

The potential benefits of hypothermia on outcome in patients after cardiac arrest is beyond the scope of this chapter and is discussed in Chap. 16. However, when intracranial hypertension does occur as a result of global cerebral hypoperfusion it is a reflection of diffuse neuronal injury. As a result, when secondary intracranial hypertension occurs in this situation, it signals widespread injury and poor neurologic outcome. We strongly discourage aggressive management of this complication in these patients.

Large Supratentorial Hemispheric Infarctions

While large supratentorial cerebral hemispheric infarctions (LHI) are not common (accounting for less than 10% of all ischemic strokes), they are among the most disabling and deadly. As a result, physicians involved with the management of these patients must be equipped with a contemporary management strategy to minimize disability and mortality in patients in whom survival is desired as the appropriate medical care focus in keeping with the patient's life philosophy. LHI defines a group of patients with disabling strokes, variable degrees of collateral circulation, and brain swelling and life-threatening deterioration from brain herniations and intracranial hypertension.

In addition to the usual priorities of general systemic care (e.g., respiratory, cardiovascular, and nutritional), and general stroke care (e.g., blood sugar control, fever management, and deep vein thrombosis prophylaxis), patients who suffer a LHI should receive thoughtful application of medical treatments and monitoring for optimizing brain perfusion (or avoiding cerebral hypoperfusion), minimizing brain swelling, and limiting brain tissue shifts. There should be early discussion with the patient, family, and surrogate regarding the patient's life priorities as they may apply to practical life-and-death decision making and procedures in the context of the disabling stroke event; a strategic monitoring plan for early detection of deterioration and brain swelling; and engagement of other professionals necessary for the timely application of treatments necessary in the case of significant worsening (e.g., neurosurgeons).

There have been a variety of clinical predictors studied and identified to correlate with fatal outcome from LHI. Some of these factors include high National Institutes of Health Stroke Scale scores, early drowsiness, and early nausea and vomiting.71–75 The various identified prognostic factors are generally associated with larger infarctions, and not surprisingly, CT and magnetic resonance imaging (MRI) analyses have confirmed a correlation between infarction volume and outcome from supratentorial infarctions.76,77 While new evolving approaches to predicting brain swelling in patients with LHI are exciting, and our wisdom on their best application will evolve over time, at this point all acute patients with LHI should be considered at risk for severe, life-threatening deterioration. This is supported by the preliminary results of the recently completed HeADDFIRST study, a prospective randomized pilot clinical trial on LHI and surgical decompression. In that study, 65% of the registered patients with at least complete MCA territory infarction (based on acute clinical and CT imaging criteria) developed life-threatening brain swelling and tissue shifts (≥7 mm of anteroseptal shift or ≥4 mm of pineal shift from midline) within 96 hours of stroke onset.78

While patients with acute ischemic stroke are a heterogeneous group with variable baseline blood pressure and stroke mechanisms, those patients with LHI more likely have large-vessel narrowing or occlusion, autoregulatory dysfunction, and/or are vitally dependent on collateral circulation. At this point in our understanding, blood pressure lowering should only be done with great reluctance in patients with LHI, with clearly prioritized goals, thoughtful agent selection (to be discussed), and vigilant monitoring to avoid overtreatment. Depending on the extent of brain swelling and the degree of blood pressure lowering desired, it may be rational and appropriate to consider parenchymal ICP or CBF monitoring to avoid exacerbating regional cerebral hypoperfusion. In most cases, we recommend maintenance of the blood pressure at least in the high-normal range with LHI in order to maintain collateral perfusion, since cerebral edema progressively challenges this vital brain-preserving source of cerebral blood flow.

It has been shown that the majority of patients who deteriorate from LHI do not have important ICP elevations or cerebral hypoperfusion as an early contributing factor to their worsening.35 Their clinical deterioration is mainly due to BTD from evolving brain pressure differentials caused by regional cerebral edema (Fig. 65-11). Indiscriminant administration of mannitol or contralateral ventricular drain insertion and CSF drainage (the ipsilateral ventricle is usually collapsed) can lead to accentuation of the pressure differentials that drive BTD and augment the clinical worsening. However, early ICP elevation, when it does occur with LHI, stratifies the patient to higher risk of death from brain swelling, and younger patients are at higher risk for such early elevations.35,79 It has not been shown that ICP-focused management (whether or not under the guidance of ICP monitoring) improves the outcome in LHI patients, but it has not been well-studied with appropriate standardized medical treatment protocols. In fact, some of the more widely quoted studies on LHI with aggressive ICP lowering–focused treatment strategies report some of the most dismal outcomes for this disease.80,81

We use ICP monitoring in young patients with LHI who already have shown evidence of significant regional brain swelling and compression of ipsilateral cerebrospinal fluid spaces. These patients have declining intracranial compliance and are at risk for ICP plateau waves. The presence of early ICP elevation in the young patient should put everyone on the alert for the escalating risk of death in that individual patient, and this can be factored into decisions regarding treatment escalation and the possible application and timing of surgical decompression. In addition, ICP monitoring in such patients may assist with strategies (positioning and medical) to improve intracranial compliance, and attempt to avoid catastrophic cerebral hypoperfusion during transport and various nursing maneuvers. When ICP monitoring is employed, we recommend using an ipsilateral (to the infarction) parenchymal monitor, since it will be the region of greatest ICP elevation. The ipsilateral placement is because the ICP will be maximal in the region of dominant brain swelling, and the ventricle ipsilateral to the infarction collapses as swelling progresses, with the various disadvantages mentioned earlier in this section.

When ICP is elevated, medical management alone carries a high mortality. However, with respect to ICP reduction, the application of conventional ICP-lowering strategies can be very helpful early in the course of management. These include: optimizing head and body positioning, avoidance of behaviors that elevate ICP in patients with poor cerebral compliance (fighting the ventilator, agitation, and seizures), tight fever control, hyperventilation, and the administration of mannitol. We rarely administer multidose regimens of mannitol in LHI patients, because its misapplication and overuse has hypothetical dangers in these patients. When ICP elevations are considered important to medically treat in patients with LHI, use of hypertonic saline can also be an effective strategy.

The theme of fluid management in patients with LHI should be avoidance of volume depletion. There is no role for dehydration as a management strategy to limit the development of brain edema. Isotonic fluids (0.9% saline) are generally recommended without clear scientific evidence for their unique advantage for patients with LHI. When patients have been exposed to hyperosomolar treatments, the use of only isotonic fluids is critical, as the hyperosmolar brain may be at risk for worsened cerebral edema when exposed to hypotonic fluids. There is some evidence of a possible benefit of albumin infusions in experimental animals with LHI from MCA occlusion, but this has not been widely applied in humans.82,83

The potential for surgical decompression as a method to limit infarct volume and mortality from brain swelling after stroke has been demonstrated in experimental animals and with more recently promising human work.78,81,84−86 Unfortunately, the selection of patients for this procedure and its best timing are still unclear; the decision making for this step requires a delicate balance between the patient's medical condition, the pace and anticipated severity of clinical progression, the prognosis, premorbid patient wishes, and various ethical and psychosocial issues.