Critical Care Management of Increased Intracranial Pressure

ANALYTIC REVIEWS

Critical Care Management of Increased Intracranial Pressure
Stephan A. Mayer, MD*² and Ji Y. Chong, MD*
Mayer SA, Chong JY. Critical care management of increased intracranial pressure. J Intensive Care Med 2002;17:55±67.

Increased intracranial pressure (ICP) is a pathologic state common to a variety of serious neurologic conditions, all of which are characterized by the addition of volume to the intracranial vault. Hence all ICP therapies are directed toward reducing intracranial volume. Elevated ICP can lead to brain damage or death by two principle mechanisms: (1) global hypoxic-ischemic injury, which results from reduction of cerebral perfusion pressure (CPP) and cerebral blood ¯ow, and (2) mechanical compression, displacement, and herniation of brain tissue, which results from mass effect associated with compartmentalized ICP gradients. In unmonitored patients with acute neurologic deterioration, head elevation (30 degrees), hyperventilation (pCO2 26±30 mmHg), and mannitol (1.0±1.5 g/kg) can lower ICP within minutes. Fluid-coupled ventricular catheters and intraparenchymal pressure transducers are the most accurate and reliable devices for measuring ICP in the intensive care unit (ICU) setting. In a monitored patient, treatment of critical ICP elevation (>20 mmHg) should proceed in the following steps: (1) consideration of repeat computed tomography (CT) scanning or consideration of de®nitive neurosurgical intervention, (2) intravenous sedation to attain a quiet, motionless state, (3) optimization of CPP to levels between 70 and 110 mmHg, (4) osmotherapy with mannitol or hypertonic saline, (5) hyperventilation (pCO2 26±30 mmHg), (6) high-dose pentobarbital therapy, and (7) systemic cooling to attain moderate hypothermia (32±33°C). Placement of an ICP monitor and use of a stepwise treatment algorithm are both essential for managing ICP effectively in the ICU setting.

Increased intracranial pressure (ICP) can result from a number of insults to the brain, including traumatic brain injury (TBI), stroke, encephalitis, neoplasms, and abscesses (Table 1). The fundamental abnormality common to these diverse disease states is an increase in intracranial volume. Accordingly, all treatments for elevated ICP work by reducing intracranial volume. Prompt recognition and treatment of elevated ICP is essential because sustained elevated ICP can cause brain damage or be rapidly fatal.

Pathophysiology
Intracranial Compliance. The Monroe±Kellie doctrine dictates that the cranial vault is a ®xed space that contains three compartments: blood, cerebrospinal ¯uid (CSF), and brain tissue. In the average adult, the brain volume is 1400 ml, the blood volume is 150 ml, and the CSF volume is 150 ml. CSF is produced by the choroid plexus in the ventricles at a rate of approximately 20 ml/hr, and drains into the venous system via the arachnoid villi and granulations [1]. This out¯ow is normally of low resistance; hence jugular venous pressure is the chief determinant of ICP in healthy patients. Normal ICP ranges from 50 to 200 mmH2O or 3±15 mmHg. In routine intensive care unit (ICU) practice, the goal of ICP management is to maintain levels below 20 mmHg. In pathologic states characterized by increased ICP (Table 1), additional volume is added to the intracranial compartment. This can result from the addition of an extrinsic mass lesion or from an increase in the volume of CSF (hydrocephalus), brain tissue (cytotoxic edema), or blood (vasogenic edema). To maintain ICP within normal limits, these increases in intracranial volume are initially counterbalanced by volume reductions in the other compartments. CSF is displaced through the foramen magnum into the paraspinal space, blood is displaced from the intracranial to the extracranial venous system, and the brain parenchyma is compressed. After these mechanisms are exhausted,
Copyright Ó 2002 Blackwell Science, Inc. 55

From the Division of Critical Care Neurology, Departments of *Neurology and ²Neurosurgery, College of Physicians and Surgeons, Columbia University, New York, NY. Received Jul 30, 2001, and in revised form Aug 24, 2001. Accepted for publication Aug 28, 2001. Address correspondence to Dr. Mayer, Neurological Institute, 710 West 168th St., Unit 39, New York, NY 10032, or e-mail: [email protected].

56

Journal of Intensive Care Medicine Vol 17 No 2 March/April 2002

Table 1. Conditions Associated with Increased ICP Intracranial mass lesions Subdural hematoma Epidural hematoma Brain tumor Cerebral abscess Intracerebral hemorrhage Increased brain volume (cytotoxic edema) Cerebral infarction Global hypoxia-ischemia Reye's syndrome Acute hyponatremia Increased blood and brain volume (vasogenic edema) Hepatic encephalopathy Traumatic brain injury Meningitis Encephalitis Hypertensive encephalopathy Eclampsia Subarachnoid hemorrhage Dural sinus thrombosis Altitude-related cerebral edema Increased CSF volume Communicating hydrocephalus Noncommunicating hydrocephalus Choroid plexus papilloma Fig 1. Intracranial pressure-volume relationship. At point A, on the ¯at portion of the curve, the amplitude of the arterial re¯ection in the ICP waveform is small (inset), and the addition of volume leads to a small increase, in pressure (A¢). At point B, on a steeper portion of the curve, the intracranial compartment is relatively noncompliant, the amplitude of the arterial re¯ection in the ICP waveform is large (inset), and addition of the same amount of volume leads to a larger increase, in pressure (B¢). (Reprinted from Mayer SA. Management of increased intracranial pressure. In: Wijdicks EFM, Diringer MN, Bolton CF, et al. Continuum: Critical Care. Minneapolis, MN: American Academy of Neurology, 1997:47±61.)

intracranial compliance (DV/DP) falls sharply, and even small increases in intracranial volume can lead to dramatic elevations in ICP (Fig 1). The relative state of intracranial compliance can be assessed by inspection of the ICP waveform (Fig 1, insets). ICP normally increases 2±3 mmHg with each arterial pulse because of transient increases in cerebral blood volume (CBV). When intracranial compliance is reduced, intracranial pulse pressure can reach levels of 10±15 mmHg, which re¯ects loss of the ability to accommodate even small pulsatile increases in CBV. As intracranial compliance falls, the morphology of the ICP waveform also changes, in that the amplitude of the second peak (the dicrotic wave) initially equals and then exceeds the amplitude of the ®rst peak (the percussion wave) (Fig 2). Cerebral Perfusion Pressure. Cerebral perfusion pressure (CPP), de®ned as the mean arterial pressure (MAP) minus ICP, is a critical determinant of cerebral blood ¯ow (CBF) and plays an important role in ICP management. Normally CBF is ``autoregulated'' at a constant level over a wide range of CPPs (from 50 to 150 mmHg) (Fig 3). Pressure autoregulation of this type is mediated by changes in arteriolar diameter and cerebrovascular resistance. The autoregulatory curve is shifted to the left in children and shifted to the right in patients with chronic hypertension. In pathologic states with impaired autoregulation, such as TBI and

Fig 2. ICP waveform in conditions of normal (top) and abnormal (bottom) intracranial compliance. (Reprinted from Chestnut RM, Marshall LF. Treatment of abnormal intracranial pressure. Neurosurg Clin N Am 1991;2:267± 284.)

subarachnoid hemorrhage, CBF may approximate a linear relationship with CPP, which creates a smaller range of optimal CPP (Fig 2). Reduction of CPP below the lower limit of autoregulation can lead to ischemia [2], whereas CPP elevation above the upper limit of autoregulation can be associated with hyperemia, exacerbation of vasogenic edema, and increased ICP [3]. Although the optimal CPP for any particular patient may vary, as a rule of thumb

Mayer and Chong: Management of ICP

57

Fig 3. Cerebral autoregulation curve. In the normal relationship (solid line), with CBF held constant across a wide range of CPP (50±150 mmHg). In disease states (e.g., vasospasm, ischemia, intracranial mass lesion), cerebral blood ¯ow may become pressure passive (dotted line). (Reprinted from Mayer SA. Management of increased intracranial pressure. In: Wijdicks EFM, Diringer MN, Bolton CF, et al. Continuum: Critical Care. Minneapolis, MN: American Academy of Neurology, 1997:47±61.)

Fig 4. Pathologic ICP elevations. (A) Lundberg A (plateau) waves. (B) Lundberg B waves. (Reprinted from Mayer SA. Management of increased intracranial pressure. In: Wijdicks EFM, Diringer MN, Bolton CF, et al. Continuum: Critical Care. Minneapolis, MN: American Academy of Neurology, 1997:47±61.)

CPP should be maintained above 70 mmHg to avert ischemia, and below 110 mmHg to avoid breakthrough hyperperfusion. To accurately measure CPP in the ICU, the pressure transducer used to measure MAP must be set at head level [4]. CBF also depends upon PaCO2 and PaO2 levels. In general, the cerebral vessels are less responsive to changes in PaO2 than to those in PaCO2. Arteriolar diameter and CBF progressively increase as PaCO2 rises from 20 to 80 mmHg, whereas hypoxemia leads to vasodilation and increased CBF only when PaO2 falls below 50 mmHg [1]. Pathologic ICP Waves. Patients with reduced intracranial compliance and elevated ICP may develop pathologic ICP waves (Fig 4). Lundberg A waves (plateau waves) are dangerous elevations in ICP [5]. They occur suddenly, can reach levels of 50±100 mmHg, and can last from minutes to hours. Plateau waves are characteristically associated with ``mirror'' reductions in CPP (Fig 5). When severe, plateau waves are associated with reduced CPP and CBF, leading to global hypoxic-ischemic damage. Lundberg B waves are of lesser amplitude (5±20 mmHg) and of shorter duration (1±5 minutes) than A waves. Although they are not directly harmful, B waves are a useful marker of abnormal intracranial compliance. Both of these waves characteristically end with a surge in systemic blood pressure and ICP, known as a termination spike [5,6]. A vasodilatory cascade model has been proposed to explain the pathogenesis of pathologic ICP waves [6,7]. According to this model, pathologic A and B waves occur because CPP is inadequate. The

process begins with a reduction in CPP, which results from a drop in MAP or a surge in ICP [6]. To maintain CBF, the cerebral vasculature dilates and CBV increases. This adds volume to the intracranial compartment, resulting in an increase in ICP and further reduction of CPP. This initiates a vicious cycle in which cerebral vasodilation continues until it is maximal, at which point a ``plateau'' is reached at a new level of increased CBV and ICP and decreased CPP and CBF. The plateau ends once CBF is inadequate to maintain tissue oxygenation and ischemia develops. This results in a re¯ex systemic pressor response mediated by a surge in systemic vascular resistance. MAP climbs and CPP is restored (the termination spike), which allows the

Fig 5. Plateau waves followed by sustained ICP elevation in a patient with traumatic brain injury. Early in the course of the recording, plateau waves exceeding 90 mmHg are associated with ``mirror'' reductions in CPP below 50 mmHg. At the end of the recording, ICP remains elevated between 40 and 60 mmHg, MAP falls below 75 mmHg, and CPP drops to 20 mmHg. At this point the patient became clinically brain dead.

58

Journal of Intensive Care Medicine Vol 17 No 2 March/April 2002

cerebral vessels to return to normal caliber, restoring CBV and ICP to normal levels.

Signs of Increased ICP and Herniation
The clinical manifestations of increased ICP are well known, but are notoriously unreliable (Table 2). A depressed level of consciousness and re¯ex hypertension, the two most consistent signs, both re¯ect the effects of globally reduced CBF. However, many patients have multiple reasons for a depressed level of consciousness, and in some patients with signi®cant shift and mass effect, ICP may be normal. Comatose patients with intracerebral hemorrhage [8] and middle cerebral artery territory infarction [9], for instance, may have ICP levels that vary from normal to highly elevated, and brain stem herniation can occur in the absence of elevated ICP. Cushing's triad (hypertension, bradycardia, and irregular respiration), which was originally described in response to elevated ICP, can also result from brain stem herniation. Because of the poor correlation between clinical signs and ICP, the only way to properly diagnose increased ICP is to directly measure it. It is important to differentiate clinical signs of increased ICP from signs of cerebral herniation (Table 3). Brain tissue displacement and herniation occur when compartmentalized mass effect leads to ICP gradients [10,11]. Patients with intracranial mass
Table 2. Clinical Signs of Increased ICP Signs which are almost always present Depressed level of consciousness (lethargy, stupor, oma) Hypertension, with or without bradycardia Symptoms and signs which are sometimes present Headache Vomiting Papilledema Sixth cranial nerve palsies

lesions may have elevated ICP, brain tissue shifting, or both. Herniation is often rapidly fatal, but can be reversed by reducing mass effect related to compartmentalized ICP gradients with treatments such as mannitol, hypertonic saline, and hyperventilation.

ICP Monitoring
Indications. Invasive monitoring of ICP is generally indicated in patients who meet all three of the following criteria: 1. The patient is suspected to be at risk for elevated ICP. 2. The patient is comatose (Glasgow coma scale score £ 8). 3. The prognosis is such that aggressive ICU treatment is indicated. Suspicion of increased ICP is usually based on clinical signs (Tables 2 and 3) and the results of a computed tomography (CT) scan showing signi®cant intracranial mass effect with midline shift or effacement of the basal cisterns. However, in comatose patients with TBI, intracranial hypertension occurs in approximately 10% of patients with normal CT scans; this risk is even higher in patients more than 40 years old, with motor posturing, or with hypotension (systolic blood pressure < 90 mmHg) [12]. If a patient is awake and can follow commands, it is unlikely that ICP is dangerously elevated [13], and the bene®ts of ventricular drainage or ICP monitoring probably do not outweigh the risks. Careful monitoring of mental status in an ICU will usually suf®ce in these cases. Invasive ICP monitoring devices. Empiric therapy for increased ICP (i.e., standing doses of mannitol) without invasive monitoring is a distressingly common practice. This approach is unsatisfactory

Table 3. Herniation Syndromes Type Uncal (lateral transtentorial) Central transtentorial Clinical hallmark Ipsilateral CN 3 palsy Contralateral motor posturing Coma with progression from bilateral decorticate to decerbrate posturing Rostral-caudal loss of brain stem re¯exes Coma with asymmetric (contralateral > ipsilateral) motor posturing Sudden progression to coma with bilateral motor posturing in a patient with cerebellar signs Causes Temporal lobe mass lesion Diffuse cerebral edema Acute hydrocephalus Convexity (frontal or parietal) mass lesion Cerebellar mass lesion

Subfalcine Cerebellar (upward or downward)

Mayer and Chong: Management of ICP

59

because most ICP treatments are effective for a short time only, may lose their ef®cacy with prolonged use, and have side effects. Optimally therapy should be given when ICP is high, and withheld when it is normal. Only an invasive ICP monitor makes this possible. The four main types of invasive ICP monitors used in standard ICU practice are listed below, in general order of their accuracy and reliability (Fig 6).
INTRAVENTRICULAR CATHETERS. These devices directly connect the intracranial space to an external pressure transducer via saline-®lled tubing. The bedside pressure transducer must be positioned at the level of the foramen of Monroe (external auditory meatus) to accurately re¯ect ICP. The catheter is usually connected to both a pressure transducer and an external drainage system via a three-way stopcock. The system can then be set for continuous ICP monitoring with intermittent CSF drainage or continuous drainage with intermittent ICP measurement. The major advantage to intraventricular catheters (IVCs) is that they allow treatment of increased ICP via drainage of CSF. The main disadvantage is the high risk of infection (ventriculitis or meningitis), which occurs in 10±20% of patients and increases dramatically after 5 days [14]. Most neurointensivists administer prophylactic antibiotics with gram-positive coverage, such as oxacillin 1±2 g every 6 hours, to minimize this risk. In a clinical trial of 228 IVC patients, prolonged therapy with ampicillin sulbactam in the ICU signi®cantly reduced the frequency of CSF infection compared to patients given perioperative treatment only (3% versus 11%) [15].

INTRAPARENCHYMAL PRESSURE TRANSDUCERS. The pressure transducer in these disposable devices is incorporated into the tip of a thin ®beroptic cable (the Camino device) or within a strain-gauge microsensor at the tip of a ¯exible catheter (the Codman device). These catheters can be placed into either the brain parenchyma or the ventricle via a small burr hole and screw [16,17]. With intraparenchymal placement, the infection rate is exceedingly low (approximately 1%) [18]. When combined with a ventricular catheter, the system allows simultaneous CSF drainage and continuous ICP measurement. These devices only need to be calibrated once prior to insertion, and the accuracy of ICP measurements is generally superior to those provided by subarachnoid bolts or epidural transducers [19]. A new version of the Codman monitor also provides measurements of brain temperature. A third parenchymal monitor recently approved by U.S. Food and Drug Administration (FDA) (the Spielberg device) features a small air-®lled balloon at the tip of a ¯exible catheter; it has the advantage of providing measurements of intracranial compliance (calculated as a pressure/volume index) as well as ICP [20]. SUBARACHNOID BOLTS. This is another ¯uid-coupled system which connects the intracranial space to an external transducer at the bedside via saline-®lled tubing [21]. The subarachnoid bolt is actually a hollow screw that is inserted via a burr hole. The dura at the base of the bolt is perforated with a spinal needle, allowing the subarachnoid CSF to ®ll the bolt. Pressure tubing is then connected to establish communication with a pressure monitoring system. Although the infection risk is low, these devices are prone to error, including underestimation of ICP, screw displacement, and occlusion by debris [22]. EPIDURAL TRANSDUCERS. These devices (the Gaeltec device) are inserted deep into the inner table of the skull and super®cial to the dura [23]. In most of these devices, pressure is transduced by an optical sensor. They have a low infection rate (approximately 1%) [17], but are prone to malfunction, displacement, and baseline drift that can exceed 5±10 mmHg after more than a few days of use. Much of the inaccuracy results from having the relatively inelastic dura between the sensor tip and the subarachnoid space.

Fig 6. ICP monitoring devices. (A) Intraparenchymal ®beroptic probe or microsensor; (B) intraventricular catheter; (C) epidural transducer; (D) subarachnoid bolt. (Reprinted from Mayer SA. Management of increased intracranial pressure. In: Wijdicks EFM, Diringer MN, Bolton CF, et al. Continuum: Critical Care. Minneapolis, MN: American Academy of Neurology, 1997:55.)

Noninvasive ICP Monitoring. At present there is no noninvasive method that can provide accurate continuous on-line measurement of ICP. However, transcranial Doppler (TCD) ultrasonography, which measures the velocity of blood ¯ow in the

60

Journal of Intensive Care Medicine Vol 17 No 2 March/April 2002

basal cerebral arteries, shows characteristic changes with increasing ICP [24]. As CPP falls, diastolic velocity decreases and pulsatility increases, re¯ecting increased distal vascular resistance to ¯ow. Though this ®nding is speci®c for severe intracranial hypertension, TCD is not sensitive to mild to moderate ICP elevations. Lateralized asymmetries in TCD pulsatility correlate with lesion volume in intracerebral hemorrhage, and are believed to re¯ect compartmentalized ICP gradients [25]. Promising new applications using ultrasound technology to estimate ICP noninvasively have been described, but have not yet been validated in the clinical setting [26±28]. Adjuncts to ICP Monitoring. Several new modalities have recently been introduced that can provide additional information regarding the adequacy of cerebral perfusion and the extent of injury in patients undergoing ICP monitoring.
BRAIN TISSUE PO2 (PBTO2) MONITORING. SjvO2 monitoring assesses the adequacy of global cerebral oxygen delivery, whereas PbtO2 monitoring measures regional oxygen tension. Both emerging technologies provide continuous information regarding the adequacy of CPP and CBF at the tissue level [29±31]. SjvO2 is measured with a 5-French ®beroptic oxygen saturation catheter placed retrograde in the internal jugular vein so that the tip is positioned in the jugular bulb; PbtO2 is measured with a miniaturized Clark electrode embedded in the tip of a thin catheter inserted 3±4 cm into the cerebral white matter (the Licox or Neurotrend device). Both techniques can detect inadequate CBF (i.e., ischemia), which may occur even in patients with relatively normal CPP [32], and excessive CBF (i.e., hyperemia), which can aggravate ICP related to vasogenic edema and breakthrough of autoregulation [33]. Accordingly these monitors can be used to optimize therapy: mannitol and vasopressor infusion reduce ICP and improve cerebral oxygenation when SjvO2 or PbtO2 values fall below critical levels [30±32], whereas hyperventilation reduces ICP and tissue oxygenation when it is supranormal due to relative hyperperfusion [31] (Table 4). The depth and duration of ischemia detected by either device is highly correlated with poor clinical outcome in patients with severe TBI [30,31]. Since neither monitor alone can detect all episodes of ischemia [34], selection of which type to use depends on the speci®c clinical circumstances at hand. PbtO2 monitoring is generally easier to use, and is most desirable when detection of ischemia in a speci®c brain region is the JUGULAR VENOUS OXYGEN SATURATION (SJVO2) AND

Table 4. Normal and Critical Values for SjvO2 and PbtO2 Monitors SjvO2 (%) Normal valuea Critical upper limit (indicates hyperemia) Critical lower limit (indicates ischemia) Average value in severe TBIa 62 80 50 73 ‹ 10 PbtO2 (mm Hg) 37 NA 8 32 ‹ 19

NA ˆ data not available. Interpretation of critical values assumes normal FiO2 (£40%) and hematocrit. a Data from reference 34.

predominant concern, whereas SjvO2 is less in¯uenced by high FiO2 levels and hence may be a more reliable measure of relative hyperperfusion [34].
CEREBRAL MICRODIALYSIS. This adjunctive monitoring technique is labor intensive and has yet to gain widespread acceptance. A probe is placed through the skull and levels of different substances may be measured using high-performance liquid chromatography (HPLC). Lactate, glutamate, and more recently extracellular potassium have been measured using this microdialysis. These levels correlate with cerebral ischemia and poor outcome [35]. MULTIMODAL MONITORING. Finally, a combination of multiple monitoring techniques may soon be possible. Various prototypes are in development that can allow simultaneous monitoring of ICP, PbtO2, PbtCO2, pH, brain temperature, laser Doppler ¯ow, and even microdialysis.

Management of Increased ICP
General Care Issues. Proper management of all critically ill brain-injured patients begins with general care issues designed to optimize oxygenation and cerebral blood ¯ow and to minimize factors that can aggravate neuronal injury or trigger ICP elevations. The following guidelines should be followed in all patients at risk for increased ICP:
PATIENT POSITIONING. As long as the patient is not hypotensive (mean blood pressure < 60 mmHg), the head of the bed should be elevated to 30 degrees and a straight head position should be maintained. Head elevation to 30 degrees reduces ICP by reducing jugular and cerebral venous pressure and enhancing venous out¯ow, without signi®cantly lowering CPP, CBF, or cardiac output [36,37]. Some have recommended a head ¯at position to prevent any reductions of CPP that

Mayer and Chong: Management of ICP

61

may occur with head elevation, which may be reasonable in selected cases [38]. However, head elevation in excess of 45 degrees should generally be avoided because paradoxical increases in ICP can occur in response to excessive CPP reduction [39]. Sharp head angulation should also be avoided, as it may cause jugular venous compression, increased venous backpressure, and increased ICP.
FLUID MANAGEMENT. Only isotonic ¯uids, such as 0.9% (normal) saline or lactated Ringer's solution, should be used in patients at risk for elevated ICP. Cerebral edema is generally understood to result from the creation of ``idiogenic osmoles'' which draw water down an osmolar gradient into injured brain tissue. Hypotonic ¯uids such as 5% dextrose or 0.45% (half-normal) saline should be strictly avoided because the free water contained in these ¯uids can aggravate cerebral edema and increase ICP [40]. Systemic hypo-osmolality (<280 mOsm/L) should be aggressively treated with mannitol or 3% hypertonic saline. The traditional practice of restricting total ¯uid intake (dehydration therapy), with the goal of reducing the extracellular ¯uid volume, has not been shown to signi®cantly impact brain water content or ICP. In fact, hypovolemia may lead to inadequate CPP and a consequent increase in ICP [41]. Patients with elevated ICP should have a central venous line placed to monitor central venous pressure; this is particularly critical for patients treated with mannitol or hypertonic saline. As a general rule, a central venous pressure greater than 5 mmHg and equal to slightly positive daily net ¯uid balance should be maintained by increasing the rate of infusion of isotonic crystalloid, administering 0.9% saline or 5% albumin ¯uid boluses, or transfusing blood to maintain hematocrit at greater than 24% [42]. Use of 1.5±3% hypertonic saline as a maintenance intravenous ¯uid for patients with cerebral edema is gaining popularity in the neurocritical care community. Though support for this practice to date is inconclusive, adequate clinical trials have yet to be performed [43]. A small study of pediatric TBI patients compared the use of 1.6% saline with lactated Ringer's solution and found no differences in ICP or CPP, though the hypertonic saline group required fewer interventions to lower ICP [44]. Lack of uniform concentrations and poor de®nition of dose-response relationships has limited the widespread use of hypertonic saline to date. TEMPERATURE MANAGEMENT. Fevers should be treated aggressively. Temperature elevations increase ICP by increasing cerebral metabolism and blood

¯ow, and have been shown to exacerbate hypoxicischemic neuronal injury in experimental animals [45]. As a general standard, acetaminophen and cooling blankets should be given to all patients with sustained fevers in excess of 38.3°C (101.0°F), but evidence for their ef®cacy in neurologic patients is scant [46,47]. Endovascular cooling with the use of closed-circuit water-circulating intravenous catheters is a promising new approach that is currently under development. Recent studies suggest that indomethacin may be the ideal antipyretic to use in patients with increased ICP. Indomethacin has been shown to decrease CBF and ICP in animal models and patients with TBI [48]. The mechanism of action is not known, but may involve vasoconstriction of cerebral vessels and inhibition of prostaglandin synthesis [48].
SEIZURE PROPHYLAXIS. Seizures can lead to profound elevations of CBF, CBV, and ICP, even in patients who are sedated or paralyzed [49]. This is secondary to the increased cerebral metabolic demand that occurs with seizures. Intravenous fosphenytoin (15±20 mg/kg loading dose, 3±5 mg/kg/day) is the preferred agent for seizure prophylaxis while in the ICU. STEROIDS. Dexamethasone and other steroids should not be used as a standard treatment for ICP because they are ineffective against cytotoxic edema [50]. There is generally no role for steroids in the treatment of mass effect related to cerebral infarction [51], intracerebral hemorrhage [52], or TBI [53]. By contrast, vasogenic edema related to neoplasm or abscess is steroid responsive, and dexamethasone 4±20 mg every 6 hours can lead to dramatic reductions in lesion volume [54].

Emergent Treatment of Increased ICP in an Unmonitored Patient. Emergency measures for ICP control (Table 5) are appropriate for comatose patients who present acutely with clinical signs of increased ICP or herniation. These measures are

Table 5. Emergency Measures for ICP Reduction 1. Elevate head of bed 15±30 degrees. 2. Normal saline (0.9%) at 80±100 cc/hr (avoid hypotonic ¯uids). 3. Intubate and hyperventilate (target pCO2 26±30 mmHg). 4. Mannitol 20% 1±1.5 g/kg via rapid intravenous infusion. 5. Foley catheter. 6. CT scan and urgent neurosurgical consultation.

62

Journal of Intensive Care Medicine Vol 17 No 2 March/April 2002

designed to lower ICP as quickly and effectively as possible in order to ``buy time'' before a de®nitive neurosurgical procedure (craniotomy, ventriculostomy, or placement of an ICP monitor) can be performed. Aggressive hyperventilation and mannitol therapy are the cornerstones of this type of intervention. Stepwise Treatment Protocol for Elevated ICP in a Monitored Patient. The primary goal of ICP management in a monitored patient is to maintain ICP below 20 mmHg and CPP above 70 mmHg. Contemporary ICP management has changed in recent years in two important aspects: CPP management (in addition to ICP control) has become increasingly emphasized, and the potential for overzealous hyperventilation to cause excessive vasoconstriction and aggravate ischemia has become increasingly recognized. The stepwise protocol presented here for managing ICP in a monitored patient (Table 6) re¯ects these considerations. More than one step may be instituted simultaneously, but only after all of the preceding steps have been addressed. Likewise, ICP therapy should be withdrawn in a similar stepwise fashion (Fig 7). This algorithm should be initiated any time ICP remains greater than 20 mmHg for more than 10 minutes. Though we favor the protocol presented below because of its simplicity, other treatment algorithms may be equally effective. Use of a standardized, evidence-based protocol for ICP management allows for ef®cient and well-coordinated treatment of patients within an institution, and can improve patient outcomes. In a retrospective study of patients with TBI, patients treated with a standardized algorithm of instituting and weaning ICP therapies required fewer interventions and had a
Table 6. Stepwise Treatment Protocol for Elevated ICPa in a Monitored Patient 1. Surgical decompression. Consider repeat CT scanning and de®nitive surgical intervention or ventricular drainage. 2. Sedation. Intravenous sedation to attain a motionless, quiet state. 3. CPP optimization. Pressor infusion if CPP is less than 70 mmHg, or reduction of blood pressure if CPP is greater than 110 mmHg. 4. Osmotherapy. Mannitol 0.25±1.0 g/kg intravenously (repeat every 1±6 hours as needed). 5. Hyperventilation. Target pCO2 levels of 26±30 mmHg. 6. High-dose pentobarbital therapy. Load with 5±20 mg/kg, infuse 1±4 mg/kg/hr. 7. Hypothermia. Cool core body temperature to 32±33°C.
a More than 20 mmHg for more than 10 minutes. Refer to text for details.

Fig. 7. The Columbia stepwise protocol for ICP management.

shorter duration of treatment than patients treated ad hoc [55].
AGE. The ®rst consideration in the face of an acute increase in ICP should always be whether a de®nitive intervention, such as craniotomy or ventriculostomy, should be performed to remove volume or decompress the skull. A repeat CT scan should be considered to rule out reaccumulation of an intracranial hemorrhage or worsening hydrocephalus. If a ventricular catheter is in place, the system should be opened to drainage and 5±10 ml of CSF removed. The option of some de®nitive surgical intervention should be continuously evaluated as additional steps to control ICP are added. Controlled lumbar CSF drainage (5±20 ml/hr) has been reported to reduce ICP and increase CPP in patients refractory to medical therapy and ventricular drainage alone [56]. However, this intervention is feasible only if the basal cisterns are open on CT, and even in these cases, transtentorial herniation remains a risk. Decompressive hemicraniectomy is becoming increasingly used as an intervention of last resort for patients who might otherwise require pentobarbital or hypothermia. Wide cranial decompression can de®nitively control ICP and reverse brain stem herniation, and has been reported to be effective for treating massive cerebral infarction [57], encephalitis [58], head trauma [59], and intracerebral hemorrhage [60] in nonrandomized studies. Complications of hemicraniectomy can include CSF leakage, local wound infection or meningitis, intracranial bleeding, and late hydrocephalus. STEP 2: SEDATION. Sedation is often overlooked as a key factor in ICP control. In patients with reduced intracranial compliance, ®ghting against physical restraints or ``bucking'' the ventilator can increase STEP 1: SURGICAL DECOMPRESSION OR CSF DRAIN-

Mayer and Chong: Management of ICP

63

ICP by elevating intrathoracic and jugular venous pressure. Arterial hypertension associated with agitation may further increase ICP if the patient is at the upper range of the autoregulatory curve. Before further measures are instituted, agitated patients with increased ICP should be sedated to the point where they are quiet and motionless (Ramsey level 5 or 6). A combination of a sedative-hypnotic and analgesic agent is usually most effective (Table 7). The preferred regimen is the combination of an opioid, such as fentanyl (1±3 lg/kg/hr) or sufentanil (0.1±0.6 lg/kg/hr), to provide analgesia and propofol (0.3±3 mg/kg/hr) for sedation. It is important to use drugs that are short acting, such that the agent may be stopped for frequent neurologic assessments throughout the day. One study showed that daily, scheduled interruption of sedation to examine patients not only reduced the length of ventilator dependence and ICU length of stay, but also decreased the need for other tests such as brain imaging and lumbar punctures to evaluate alterations in mental status [61]. Neurocritical care patients, even when comatose, can sense pain and require analgesia in addition to sedation. Therefore careful use of a low-dose continuous infusion opioid in addition to propofol or midazolam is recommended. Bolus injections of opioids, however, should be used with caution in patients with elevated ICP. Bolus infusions of the sufentanil, fentanyl, and alfentanil can transiently lower MAP and increase ICP due to autoregulatory vasodilation of cerebral vessels [62]. This effect is seen primarily in patients with intact autoregulation, but can also occur in patients with abnormal autoregulation [63]. Vasopressors may be used to avoid hypotension and possible re¯ex ICP elevation. Propofol may be the ideal sedative to use in patients with elevated ICP; besides the fact that it is

ultrashort acting, it has favorable effects on ICP and seizure activity, and may have neuroprotective properties. In a study comparing propofol to morphine in patients with severe TBI, patients treated with propofol had lower ICP values and more favorable long-term neurologic outcomes [64]. By contrast, paralysis with neuromuscular blocking agents such as vecuronium, pancuronium, or cis-atracurium is rarely necessary, and places patients at risk for prolonged paralysis due to critical illness myopathy.
STEP 3: CPP OPTIMIZATION. If CPP is less than 70 mmHg and ICP is greater than 20 mmHg, elevation of mean arterial blood pressure and CPP with a vasopressor such as dopamine or phenylephrine (Table 8) can lead to a re¯ex reduction of ICP, by eliminating cerebral vasodilation that occurs in response to inadequate perfusion (Fig 8). CPP optimization to levels well above 70 mmHg may be desirable in chronically hypertensive patients (whose autoregulatory curve is shifted to the right), or in patients with low PbtO2 or SjvO2 levels or Lundberg A and B waves (since these ®ndings generally re¯ect insuf®cient CPP). The widely accepted ``one size ®ts all'' approach to CPP management (>70 mmHg) is in all likelihood an oversimpli®cation, and efforts should be made to optimize CPP whenever patients fail standard therapy. One study that attempted to de®ne optimal CPP levels for severe TBI patients by analyzing receiver-operating curves found that a CPP of 55 mmHg was the critical threshold for poor outcome in adults [65]. It seems prudent to maintain CPP well above this level in clinical practice, however, to provide an extra margin of safety. In an uncontrolled study of TBI patients with CPP maintained above 70 mmHg with phenylephrine and

Table 7. Selected Short-Acting Intravenous Sedatives for ICP Management Agent Sedative-analgesic agents Morphine sulfate Fentanyl Sufentanil Sedative-hypnotic agents Propofol Midazolam Alkylphenol (ultrashort acting) Benzodiazepine (short acting) 0.6±6 mg/kg/hr 0.05±0.1 mg/kg/hr Opioid (sedative-hypnotic with analgesic properties) Opioid (short acting, 100 times more potent than morphine) Opioid (ultrashort acting) 2±5 mg IVP every 1±4 hours 0.5±3.0 lg/kg/hr 0.1±0.6 lg/kg/hr Pharmacology Dosage Range

Dosages are approximate and should be titrated to the patient's level of agitation and ICP. A combination of a sedative-analgesic and sedative-hypnotic agent may be more effective than the use of a single agent.

64

Journal of Intensive Care Medicine Vol 17 No 2 March/April 2002

Table 8. Selected Short-Acting Vasoactive Drugs for ICP Management Agent Labetolol Nicardipine Pharmacology Dosage Range

Blood pressure reduction a1 and b1 blocker 2±3 mg/min Calcium channel 5±15 mg/hr blocker a1 and b1 agonist 5±30 lg/kg/min (at high doses) 2±10 lg/kg/min

Blood pressure elevation Dopamine

concerns also exist regarding the potential for nitroprusside to directly dilate the cerebral vasculature and increase ICP. For these reasons, nitroprusside is best avoided in patients with elevated ICP, and should only be used if an ICP monitor is in place. Experimental and clinical studies have demonstrated that treatment of hypertension in patients with intracerebral hemorrhage does not exacerbate perilesional ischemia [66] or cause re¯ex vasodilation and increased ICP [67], as long as CPP remains within the normal range.
STEP 4: OSMOTHERAPY. If CPP is optimized, the patient is sedated, and ICP remains elevated, mannitol or hypertonic saline should be given. Mannitol, an osmotic diuretic, lowers ICP via its cerebral dehydrating effects. The effects of mannitol are biphasic. Rapid infusion immediately creates an osmotic gradient across the blood-brain barrier, which leads to movement of water from brain parenchyma to the intravascular compartment. The result is decreased brain tissue volume, and hence reduced ICP [68]. The secondary effect of mannitol results from its action as an osmotic diuretic. As mannitol is cleared by the kidneys, it leads to free water clearance and increased serum osmolality. This leads to a more prolonged intracellular dehydrating effect as water ¯ows down the osmotic gradient from the intracellular to the extracellular space. The initial dose of mannitol 20% solution is 1±1.5 g/kg, followed every 1±6 hours with doses of 0.25±1 g/kg as needed. Repeated mannitol doses should be given on the basis of ICP measurements rather than as scheduled doses, unless the goal is to establish and maintain a hyperosmolar state (300± 320 mOsm/L). The effect of a mannitol bolus on ICP begins within 10±20 minutes, reaches its peak between 20 and 60 minutes, and lasts for 4±6 hours, but this may vary widely between patients. Adverse effects of mannitol therapy include exacerbation of congestive heart failure (due to the initial intravascular volume expansion); volume contraction, hypokalemia, and profound hyperosmolality (after prolonged use); acute tubular necrosis (due to excessive hyperosmolality); and ``rebound'' increases in ICP [69]. Patients treated repeatedly with mannitol require frequent measurements of serum electrolytes and osmolality, careful recording of ¯uid input and output, and central venous pressure monitoring. Urinary volume losses should be replaced with normal (0.9%) saline to avoid volume depletion. Bolus infusion of 3%, 7.5%, 10%, or 24.3% hypertonic saline is gaining popularity as an alternative to bolus infusions of mannitol for ICP crises [43]. In an uncontrolled study of 3% saline/

Norepinephrine a1 and b1 agonist 0.03±0.6 lg/kg/min Phenylephrine a1 agonist

norepinephrine, better outcomes were obtained than in previously reported patients with strictly ICP-based management [41]. If MAP and ICP are elevated in a sedated patient, treatment of arterial hypertension can sometimes lead to a parallel reduction of ICP. If CPP is greater than 110 mmHg and ICP is greater than 20 mmHg, hypertension should be carefully treated with a short-acting agent (Table 7). However, extreme caution should be used to avoid reduction of CPP below 70 mmHg, which can trigger re¯ex cerebral vasodilation and ICP elevation. Nitroprusside can be particularly troublesome in this regard, and

Fig 8. Demonstration of CPP augmentation resulting in re¯ex ICP reduction. At approximately 9:30, the patient developed a plateau wave with ICP elevation from 20 to 40 mmHg, and CPP reduction from 100 to 70 mmHg. A dopamine infusion was started at 9:45, with nearly immediate normalization of ICP to 20 mmHg when CPP was restored to 100 mmHg.

Mayer and Chong: Management of ICP

65

acetate infusion (75±150 ml/hr) and intermittent boluses (250 ml) titrated to maintain serum sodium levels between 145 and 155 mEq/L, hypertonic saline lowered ICP related to TBI and brain tumors, but not in patients with intracerebral hemorrhage or cerebral infarction [70]. Infusion of 2±5 ml/kg of 7.5% saline or 0.5±1.0 ml/kg of 23.4% saline over 30 minutes has also been shown to lower elevated ICP and augment CPP, with an effect that lasts several hours [71,72]. A large head-to-head trial of mannitol versus hypertonic saline is warranted.
STEP 5: HYPERVENTILATION. As a general rule, the goal of hyperventilation for ICP control should be to lower pCO2 to 30 mmHg, or to 25±30 mmHg in extreme cases. The respiratory alkalosis caused by hyperventilation lowers ICP by causing cerebral vasoconstriction and reduced CBV [1]. The peak effect of hyperventilation on ICP is generally reached within 30 minutes. Over the next 1±3 hours the effect gradually diminishes, as compensatory acid-base buffering mechanisms correct the alkalosis within the central nervous system [1]. In patients with elevated ICP due to hyperemia and increased CBV, however, the effect may be prolonged, and hyperventilation may be the treatment of choice [73]. Hyperventilation should be tapered slowly over 4±6 hours because abrupt cessation may lead to vasodilation and rebound increases in ICP. Though it is postulated that prolonged severe hyperventilation (pCO2 < 25 mmHg) can actually exacerbate cerebral ischemia by causing profound vasoconstriction, the deleterious effects of excessive hyperventilation may also include more labile ICP. In a study of patients with TBI, prophylactic hyperventilation to a pCO2 of 25 mmHg resulted in poorer outcome compared with normally ventilated patients [74]. The authors hypothesized that blood vessels may become hypersensitive to changes in pCO2 after prolonged hyperventilation because CSF buffering capacity is lost. In cases of pediatric head trauma, profound hypocarbia was associated with decreased cerebral oxygen consumption and ischemia [75]. SjvO2 monitoring is a useful modality for ensuring that prolonged aggressive hyperventilation, if necessary, is not critically reducing oxygen delivery to the brain. STEP 6: PENTOBARBITAL THERAPY. High-dose barbiturate therapy, given in doses equivalent general anesthesia, can effectively lower ICP in most patients refractory to the steps outlined above [76,77]. The effect of pentobarbital is multifactorial, but most likely stems from a profound reduction of cerebral metabolism, which is coupled to reductions of CBF and CBV [78]. Pentobarbital often

causes hypotension, and usually requires the use of vasopressors to maintain CPP above 70 mmHg. Pentobarbital typically requires a loading dose of 10±20 mg/kg, given in repeated 5 mg/kg boluses, until a state of ¯accid coma with preserved pupillary reactivity is attained. Maintenance infusion is usually about 1±4 mg/kg/hr. Continuous EEG monitoring is helpful to avoid oversedation, since generally no further ICP reduction occurs once a burstsuppression pattern is attained. If ICP is normalized with pentobarbital, it is generally maintained for 24±48 hours. It can then be abruptly discontinued because its highly lipophilic nature and long halflife (90 hours) result in a gradual reduction of blood levels over several days.
STEP 7: HYPOTHERMIA. Systemic hypothermia to levels of 32±33°C can lower ICP in some patients refractory to pentobarbital [79±81]. This technique requires placement of cooling blankets under and over the patient, iced gastric lavage, and pharmacologic paralysis with vecuronium or a similar neuromuscular blocking agent to prevent shivering. Prolonged hypothermia can be dangerous because of increased risk of infectious complications, coagulopathy, and electrolyte derangements, among other hazards. Rewarming should always be done slowly, over at least a day, and passively, without active heating, to avoid rebound cerebral edema or a systemic in¯ammatory response syndrome, which can be fatal. In a small randomized controlled trial of severe TBI patients refractory to pentobarbital, mild-tomoderate hypothermia (34°C) signi®cantly reduced ICP, improved CPP, reduced CBF and cerebral metabolic rate (CMRO2), and reduced arteriojugular venous oxygen differences [79]. Survival was 50% in hypothermia patients compared to 18% in the control group (p < 0.05). Later studies by the same group reported that hypothermia was most effective for pentobarbital-refractory ICP elevations between 20 and 40 mmHg [80], and that severe TBI patients with low ICP do not bene®t from hypothermia [81]. These reports, and the negative results of a recent large National Institutes of Health (NIH)funded trial studying the effects of hypothermia as ®rst-line therapy for severe TBI [82], suggest that hypothermia is not effective as a primary form of neuroprotection for severe TBI related to diffuse axonal injury.

References
1. Ropper AH, Rockoff MA. Physiology and clinical aspects of raised intracranial pressure. In: Ropper AH, ed. Neurological

66

Journal of Intensive Care Medicine Vol 17 No 2 March/April 2002

2. 3. 4. 5. 6. 7. 8. 9. 10. 11.

12.

13. 14. 15. 16. 17. 18. 19.

20.

21. 22. 23.

and neurosurgical intensive care, 3rd ed. Philadelphia: Lippincott-Raven, 1993:11±28 Jennet WB, Harper AM, Meller JD, et al. Relationship between cerebral blood ¯ow and cerebral perfusion pressure. Br J Surg 1970;57:390±397 Shinnoj E, Strangaard S. Pathogenesis of hypertensive encephalopathy. Lancet 1973;1:461±462 Nates JL, Niggemeyer LE, Anderson MB, et al. Cerebral perfusion pressure monitoring alert! [letter]. Crit Care Med 1997;25:895±896 Lundberg N. Continuous recording and control of ventricular ¯uid pressure in neurosurgical practice. Acta Psychiatr Scand Suppl 1960;149:1±19 Rosner MJ, Becker DP. Origin and evolution of plateau waves: experimental observations and a theoretical model. J Neurosurg 1984;60:312±324 Czosnyka M, Smielewski P, Piechnik S, et al. Hemodynamic characterization of intracranial pressure plateau waves in head-injured patients. J Neurosurg 1999;91:11±19. Ropper AH, King RB. Intracranial pressure monitoring in comatose patients with cerebral hemorrhage. Arch Neurol 1984;41:725 Frank J. Large hemispheric infarction, deterioration and intracranial pressure. Neurology 1995;45:1286±1290 Wol¯a CE, Luerssen TG, Bowman RM, et al. Brain tissue pressure gradients by expanding frontal epidural mass lesion. J Neurosurg 1996;84:642±647 Sahuquillo J, Poca M-A, Arribas M, et al. Intrahemispheric supratentorial intracranial pressure gradients in head-injured patients: Are they clinically important? J Neurosurg 1999;90:16±26 Narayan RK, Kishore PR, Becker DP, et al. Intracranial pressure: To monitor or not to monitor? A review of our experience with severe head injury. J Neurosurg 1982;56:650±659 Papo I, Janny P, Caruselli G, et al. Intracranial pressure time course in primary intracerebral hemorrhage. Neurosurgery 1979;4:504±511 Mayhall CG, Archer NH, Lamb VA, et al. Ventriculostomy related infections: a prospective epidemiologic study. N Engl J Med 1984;310:553±559 Poon WS, Ng S, Wai S. CSF antibiotic prophylaxis for neurosurgical patients with ventriculostomy: a randomised study. Acta Neurochir Suppl 1998;71:146±148 Ostrup RC, Luerssen TG, Marshall LF, et al. Continuous monitoring of intracranial pressure with a miniaturized ®beroptic device. J Neurosurg 1987;67:206±209 Gopinath SP, Robertson CS, Contant CF, et al. Clinical evaluation of a miniature strain-gauge transducer for monitoring intracranial pressure. Neurosurgery 1995;36:1137±1140 Levin A. The use of a ®beroptic intracranial pressure monitor in clinical practice. Neurosurgery 1977;1:266±271 Sundbarg G, Nordstrom CH, Messeter K, et al. A comparison of intraparenchymatous and intraventricular pressure recording in clinical practice. J Neurosurg 1987;67:841±845 Yau Y-H, Piper IR, Clutton RE, et al. Experimental evaluation of the Spielberger intracranial pressure and intracranial compliance monitor. Technical note. J Neurosurg 2000;93:1072±1077 Vries JK, Becker DP, Yang HF. A subarachnoid screw for monitoring intracranial pressure. J Neurosurg 1973;39:416± 419 Miller JD, Bobo H, Kapp JP. Inaccurate pressure readings from subarachnoid bolts. Neurosurgery 1986;19:253±255 Ream AK, Silverberg GD, Corbin SD, et al. Epidural measurement of intracranial pressure. Neurosurgery 1979;5:36±43

24. Hassler W, Steinmetz H, Gawlowski J. Transcranial Doppler ultrasonography in raised intracranial pressure and in intracranial circulatory arrest. J Neurosurg 1988;68:745±751 25. Mayer SA, Thomas CE, Diamond BE. Asymmetry of intracranial hemodynamics as an indicator of mass effect in acute intracerebral hemorrhage: a transcranial Doppler study. Stroke 1996;27:1788±1792 È 26. Schmidt B, Klingelhofer J, Schwarze JJ, et al. Noninvasive prediction of intracranial pressure curves using transcranial Doppler ultrasonography and blood pressure curves. Stroke 1997;28:2465±2472 27. Pranevicius O, Bertasius K, Pranevicius K, et al. Non-invasive dynamic assessment of the elasticity of intracranial structures. Acta Neurol Scand 1992;86:512±516 28. Czosnyka M, Smielewski P, Piechnik S, et al. Critical closing pressure in cerebrovascular circulation. J Neurol Neurosurg Psychiatry 1999;66:606±611 29. Robertson C, Gopinath S, Goodman JC et al. SjvO2 monitoring in head-injured patients. J Neurotrauma 1995;12:891± 896 30. van den Brink WA, van Santbrink H, Steyerberg EW, et al. Brain oxygen tension in severe head injury. Neurosurgery 2000;46:868±878 31. Cruz J. The ®rst decade of continuous monitoring of jugular bulb oxyhemoglobin saturation: management strategies and clinical outcome. Crit Care Med 1998;26:344±351 32. Cruz J. Continuous monitoring of cerebral oxygenation in acute brain injury: injection of mannitol during hyperventilation. J Neurosurg 1990;73:725±730 33. Cormio M, Valadka AB, Robertson CS. Elevated jugular venous oxygen saturation after severe head injury. J Neurosurg 1999;90:9±15 34. Gopinath SP, Valadka AB, Uzura M, et al. Comparison of jugular venous oxygen saturation and brain tissue PO2 as monitors of cerebral ischemia after head injury. Crit Care Med 1999;27:2337±2345 35. Hutchinson P, Hutchinson D, Barr RH et al. A new cranial access device for cerebral monitoring. Br J Neurosurg 2000;14:46±48 36. Durward QJ, Amacher AL, DelMaestro RF, et al. Cerebral and cardiovascular responses to head elevation in patients with intracranial hypertension. J Neurosurg 1983;59:938± 944 37. Feldman Z, Kanter MJ, Robertson CS, et al. Effect of head elevation on intracranial pressure, cerebral perfusion pressure, and cerebral blood ¯ow in head-injured patients. J Neurosurg 1992;76:207±211 38. Rosner MJ, Coley IB. Cerebral perfusion pressure, intracranial pressure, and head elevation. J Neurosurg 1986;65:636±641 39. Moraine J, Berre J, Melot C. Is cerebral perfusion pressure a major determinant of cerebral blood ¯ow during head elevation in comatose patients with severe intracranial lesions? J Neurosurg 2000;92:606±614 40. Mayer SA. Fluid management in subarachnoid hemorrhage. Neurologist 1995;1:71±85. 41. Rosner MJ, Rosner SD, Johnson AH. Cerebral perfusion pressure: management protocol and clinical results. J Neurosurg 1995;83:949±962 42. Herbert PC, Wells G, Blajchman MA, et al. A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care. N Engl J Med 1999;340:409±417 43. Qureshi AI, Suarez JI. Use of hypertonic saline solutions in treatment of cerebral edema and intracranial hypertension. Crit Care Med 2000;28:3301±3313 44. Simma B, Burger R, Falk M, et al. A prospective, randomized and controlled study of ¯uid management in children with

Mayer and Chong: Management of ICP

67

45.

46. 47. 48.

49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61.

62. 63. 64.

severe head injury; lactated Ringer's solution versus hypertonic saline. Crit Care Med 1998;26:1265±1270 Busto R, Dietrich WD, Globus MY, et al. Small differences in intraischemic brain temperature critically determine the extent of ischemic injury. J Cereb Blood Flow Metab 1987;7:129±138 O'Donnell J, Axelrod P, Fisher C, et al. Use and effectiveness of hypothermia blankets for febrile patients in the intensive care unit. Clin Infect Dis 1996;24:1208±1213 Mayer SA, Commichau C, Scarmeas N, et al. Clinical trial of an air-circulating cooling blanket for fever control in critically-ill neurologic patients. Neurology 2001;56:292±298 Slavik R, Rhoney D. Indomethacin: a review of its cerebral blood ¯ow effects and potential use for controlling intracranial pressure in traumatic brain injury patients. Neurol Res 1999;21:491±499 Lassen NA. Control of the cerebral circulation in health and disease. Circ Res 1974;34:749±760 Fishman RA. Brain edema. N Engl J Med 1975;293:706±711 Anderson DC, Cranford RE. Corticosteroids in ischemic stroke. Stroke 1979;10:68±71 Pourgvarin H, Bhoopat TW, Viriyavejakul A, et al. Effects of dexamethasone in primary supratentorial intracerebral hemorrhage. N Engl J Med 1987;316:1229±1233 Cooper PR, Moody S, Clark WK, et al. Dexamethasone and severe head injury: a prospective double-blind trial. J Neurosurg 1979;51:307±331 Galich JM, French LA. Use of dexamethasone in the treatment of cerebral edema resulting from brain tumors and brain surgery. Am Pract 1961;12:169±174 McKinley B, Parmley C, Tonneson A. Standardized management of intracranial pressure: a preliminary clinical trial. J Trauma 1999;46:271±279 Munch EC, Bauhuf C, Horn P, et al. Therapy of malignant È intracranial hypertension by controlled lumbar cerebrospinal ¯uid drainage. Crit Care Med 2001;29:976±981 Schwab S, Steiner T, Aschoff A, et al. Early hemicraniectomy in patients with complete middle cerebral artery infarction. Stroke 1998;29:1888±1893 Schwab S, Junger E, Spranger M, et al. Craniectomy: an aggressive treatment approach in severe encephalitis. Neurology 1997;48:412±417 Guerra WK, Gaab MR, Dietz H, et al. Surgical decompression for traumatic brain swelling: indications and results. J Neurosurg 1999;90:187±196 Mayer SA, Connolly ES, Bates J, et al. Decompressive hemicraniectomy for massive cerebral hemorrhage [abstract]. Stroke 2001;32:357 Kress JP, Pohlman AS, O'Connor MF, et al. Daily interruption of sedative infusions in critically ill patients undergoing mechanical ventilation. N Engl J Med 2000; 342:1471±1477 Albanese J, Viviand X, Potie F, et al. Sufentanil, fentanyl, and alfentanil in head trauma patients: a study on cerebral hemodynamics. Crit Care Med 1999;27:407±411 de Nadal M, Ausina A, et al. Effects on intracranial pressure of fentanyl in severe head injured patients. Neurochirurgia 1998;71(suppl):10±12 Kelly D, Goodale D, Williams J, et al. Propofol in the treatment of moderate and severe head injury: a randomized, prospective double-blinded pilot trial. J Neurosurg 1999;90:1042±1051

65. Chambers I, Treadwell R, Mendelow AD. Determination of threshold levels of cerebral perfusion pressure and intracranial pressure in severe head injury by using receiveroperating characteristic curves: an observational study in 291 patients. J Neurosurg 2001;94:412±416 66. Powers WJ, Zazulia AR, Videen TO, et al. Autoregulation of cerebral blood ¯ow surrounding acute (6±22 hours) intracerebral hemorrhage. Neurology 2001;57:18±24 67. Qureshi AI, Wilson DA, Hanley DF, et al. Pharmacologic reduction of mean arterial pressure does not adversely affect regional cerebral blood ¯ow and intracranial pressure in experimental intracerebral hemorrhage. Crit Care Med 1999;27:965±971 68. Messeter K, Nordstrom CH, Sundbarg G, et al. Cerebral hemodynamics in patients with acute severe head injury. J Neurosurg 1986;64:231±237 69. Kaufmann AM, Cardoso ER. Aggravation of vasogenic cerebral edema by multiple dose mannitol. J Neurosurg 1993;44:584±589 70. Qureshi A, Suarez J, Bhardwaj A, et al. Use of hypertonic (3%) saline/acetate infusion in the treatment of cerebral edema: effect on intracranial pressure and lateral displacement of the brain. Crit Care Med 1998;26:440±446 71. Suarez J, Qureshi A, Bhardwaj A, et al. Treatment of refractory intracranial hypertension with 23.4% saline. Crit Care Med 1998;26:1118±1122 72. Hartl R, Ghajar J, Hochleuthner, et al. Hypertonic/hyperoncotic saline reliably reduces ICP in severely head-injured patients with intracranial hypertension. Acta Neurochir Suppl 1997;70:126±129 73. Obrist W, Lang®tt T, Jaggi JL, et al. Cerebral blood ¯ow and metabolism in comatose patients with acute head injury. J Neurosurg 1984;61:241±253 74. Muizelaar JP, Marmarou A, Ward JD, et al. Adverse effects of prolonged hyperventilation in patients with severe head injury: a randomized clinical trial. J Neurosurg 1991;75:731± 739 75. Skippen P, Seear M, Poskitt K, et al. Effect of hyperventilation on regional cerebral blood ¯ow in head-injured children. Crit Care Med 1997;25:1402±1409 76. Rea GL, Rockswold GL. Barbiturate therapy in uncontrolled intracranial hypertension. Neurosurgery 1983;12: 401±404 77. Eisenberg HM, Frankowski RF, Contant CF, et al. High-dose barbiturate control of elevated intracranial pressure in patients with severe head injury. J Neurosurg 1988;69:15±23 78. Cormio M, Gopinath S, Valadka A, et al. Cerebral hemodynamic effects of pentobarbital coma in head-injured patients. J Neurotrauma 1999;16:927±936 79. Shiozaki T, Sugimoto H, Taneda M, et al. Effect of mild hypothermia on uncontrollable intracranial hypertension after severe head injury. J Neurosurg 1993;79:363±368 80. Shiozaki T, Sugimoto H, Taneda M, et al. Selection of severely head injured patients for mild hypothermia therapy. J Neurosurg 1998;89:206±211 81. Shiozaki T, Kato A, Taneda M, et al. Little bene®t from mild hypothermia therapy for severely head injured patients with low intracranial pressure. J Neurosurg 1999; 91:185±191 82. Clifton GL, Miller ER, Choi SC, et al. Lack of effect of induction of hypothermia after acute brain injury. N Engl J Med 2001;344:556±563