Neural function is essential to human existence. Thus, loss of any neural element in the course of a critical illness represents a major loss to a given individual. Neurons or supporting elements may be lost in a small, virtually unnoticeable manner, perhaps manifest as cognitive or behavioral deficit, or there may be widespread selective neuronal loss or tissue infarction with more apparent and disabling deficits. Based on the notion that neural function is the essence of acceptable survival from critical illness, it is crucial for perioperative management to include considerations of neural viability and the impact and interactions of the primary diseases and therapeutics on the nervous system.
There are numerous perioperative scenarios where a patient may present with neurologic dysfunction. In a general sense, these scenarios often involve ischemia, trauma, or neuroexcitation. Each of these, as they progressively worsen, at some point typically involve a period of decreased cerebral perfusion pressure (CPP), usually resulting from elevated intracranial pressure, eventually compromising cerebral blood flow sufficiently to produce permanent neuronal loss, infarction, and possibly brain death. A variety of biochemical pathways play a major role. In this chapter we review the important physiologic factors, and intracranial pressure (ICP) considerations and therapeutic options critical to contemporary perioperative care of the patient with injury to the central nervous system (CNS).
Monitoring Neurologic Examination
The neurologic examination of the patient with an acute neurologic condition is a process that requires the physician to have a specialized anatomic and physiologic knowledge of the nervous system and the anesthesiologist should be familiar with the most basic components of the neurologic examination to identify promptly life-threatening conditions that may require urgent action. The majority of the information necessary to localize a lesion in the patient with neurologic disease can be obtained through a careful history and physical examination. Although a detailed description of neurologic examination is beyond the scope of this chapter, a rapid 5-min neurologic assessment may be sufficient to establish baseline conditions and to detect subsequent changes in neurologic status. Therefore, for the nonneurologist it is important to focus on the components of the neurologic examination described in the following sections.
Mental Status, Cranial Nerves, Motor and Sensorial Function, and Reflexes
Certain findings of the mental status examination can only be interpreted by knowing the patient’s ability to perform fundamental daily living tasks. Mental status should be reported starting with the level of alertness, followed by language function and memory. The rest of the mental status examination can be reported in any order and includes attention and concentration, memory, visual spatial recognition, praxis, calculations, etc.
Mental status is assessed by orientation to person, place, and time. General mental status is determined by assessing general information (i.e., the name of the president, obtaining a description of some current events, assessing spelling ability for simple words, and asking the patient to add or subtract serial threes or sevens). In addition, the patient can be asked to recall three objects several minutes after the examiner mentions them. Attention is tested by asking the patient to recall a series of numbers or to spell “world” backward.
Cranial nerves can be assessed as follows:
CN I (olfactory): Test each nose nare with a mild agent such as soap or tobacco.
CN II (optic): Near and far visual acuity and gross visual fields are assessed. Ophthalmoscopic examination is performed.
CN III, IV, VI (oculomotor, trochlear, abducens): Pupillary light response. Lateral and vertical gaze are assessed.
CN V (trigeminal): Assessment of touch sensation in both sides of the face is performed in all three trigeminal divisions. In addition, corneal blink reflex can also be assessed.
CN VII (facial): Symmetrical smile assessment is performed. In addition, brow wrinkling can be assessed to determine central versus peripheral seventh nerve function.
CN VIII (auditory): Ability to hear fingertips moving is assessed. Lateralizing the sound of a tuning fork placed over the mid-forehead can localize a hearing deficit.
CN IX, X (glossopharyngeal, vagus): Gag reflex is assessed with a tongue blade or tonsil tip suction device.
CN XI (accessory): Bilateral shoulder elevation (trapezius) and head-turning (sternocleidomastoid) strength is assessed.
CN XII (hypoglossal): The patient is asked to extrude his/her tongue. If there is weakness on one side of the tongue, the tongue will deviate to that side.
Motor examination is performed by assessing pronator drift of the upper extremity. A unilateral pronator drift in one arm suggests an upper motor neuron lesion affecting that arm. Hand grasps, toe and foot dorsiflexion, and the major flexors and extensors of the upper and lower extremities are assessed. In addition, individual muscles can also be assessed and graded ( Table 24.1 ).
|0 = no contraction||0 = absent|
|1 = visible muscle twitch but no movement of the joint||1 = reduced (hypoactive)|
|2 = weak contraction insufficient to overcome gravity||2 = normal|
|3 = weak contraction able to overcome gravity but no additional resistance||3 = increased (hyperactive)|
|4 = weak contraction able to overcome some resistance but no full resistance||4 = clonus|
|5 = normal|
Sensory examination involves testing the primary afferent sensory pathways with modalities such as light touch, pain and temperature, vibration and joint position, and then areas that test the ability of sensory and association cortices to interpret sensory input. These areas test sensory functions such as stereognosis, graphesthesia, point localization, etc. A brief sensory examination can be performed by stimulation with a pin on hands and feet to determine the patient’s ability to distinguish between sharp or dull sensation. The examination should be done in a repetitive manner because patients can report both stimuli as sharp bilaterally and then some 10 seconds later report to only one side as sharp, indicating a subtle deficit on the dull side. A similar maneuver is performed on the cheeks for cranial nerve V and the dorsum of the feet. It is important to note that this examination technique has been associated with the spread of infectious diseases and thus, a new pin or other sharp device should be used for each assessment with every effort made not to cause bleeding. Light touch sensation is tested in a similar manner using the touch of a finger in the place of a pin.
Coordination is assessed by finger-to-nose and heel-to-shin testing and assessment of rapid alternating movements of the hand and or foot. This aspect of the examination tests cerebellar and basal ganglia function because these structures play an important role in coordination. Deep tendon reflexes are assessed in the biceps, triceps, patella, and Achilles tendon (see Table 24.1 ). The Babinski reflexes are assessed by stroking the lateral foot. In addition to the test for meningeal irritation, the Kernig (elevation of straightened lower extremity) and Brudzinski (head flexion by examiner) tests are performed.
The neurologic examination in a comatose patient represents a different challenge. Procedures performed in evaluation of the comatose patient include determination of viable vital signs and assessment of hand drop over the head when nonphysiologic coma is suspected. Pupil size and response to light is assessed. Pinpoint pupils imply a pontine lesion or drug effect whereas large pupils can suggest a structural lesion, hypoxia, or a drug effect. Uncal herniation will produce a unilateral pupillary enlargement with ptosis and inferior–lateral position resulting from third cranial nerve lesion. Eye movements are assessed by first evaluating eye position. In destructive cerebral lesions, the eyes deviate toward the side of the lesion. Assessment of the oculocephalic (doll’s eye) reflex is performed by turning the head suddenly to one side. Eyes tend to lag behind when the patient has lethargy or is semi-comatose, but they move with the head in the awake state and when brainstem centers are impaired. Further indication of brainstem impairment is present when the corneal and gag responses are absent and when there is extensor or flexor posturing. Response to noxious stimulus is assessed peripherally by stimulating the sternum or nail beds or centrally by supraorbital pressure. The Babinski reflex is assessed.
Hourly evaluation of neurologic and mental status is recommended as part of the neuromonitoring protocol. Function of pyramidal and extrapyramidal systems, cranial nerves, cerebellum, and spinal cord whenever possible, and any trend in change of neurologic status should be recorded for every patient. In critically ill patients, however, such a complete neurologic evaluation can sometimes be unreliable or impossible owing to the use of sedatives and the need for intubation and ventilatory support as part of the medical treatment of the neurologic problem. Along with the neurologic examination, information about vital sign trends and key laboratory values should be available at all times. The Glasgow Coma Scale (GCS) ( Table 24.2 ) is used as a standardized scale for recording neurologic status in the ICU. GCS has been the standard outcome tool for neurocritical care for many years. Newer assessment tools such as the Neurological Outcome Scale for TBI (NOS-TBI) have demonstrated adequate concurrent and predictive validity as well as sensitivity to change, compared with the gold-standard outcome measure. The NOS-TBI may enhance prediction of outcome in clinical practice and measurement of outcome in TBI research.
|Type of response||Score|
Pupillary assessment is vital to determine critical neurologic conditions that may require urgent intervention. However, observer variability exists and it is subject to human error. A handheld pupilometer is a new technology that may reduce observer variability in the neurologic examination. Infrared quantitative pupillometry can produce accurate, reproducible pupillary measurements that are clearly superior to those obtained manually at the patient’s bedside by even an experienced nurse or physician. A recent study using this device reported good reliability when correlating the pupillary constriction velocity as a predictor of ICP elevation in neurosurgical patients. Neuro-ICUs are quickly adopting this type of monitoring, because evidence has shown that there is a high inter-device reliability of automated pupillometers. An important limitation of this device is that assessment is quite challenging in patients with altered mental status, in patients with periorbital or scleral edema, and in uncooperative patients. Ambient light and physiologic factors may also affect the measured pupillary characteristics.
The brain, spinal cord, cerebrospinal fluid (CSF), and blood are encased in the skull and vertebral canal, thus constituting a nearly incompressible system ( Fig. 24.1 ). In a totally incompressible system pressure would vary linearly with increased volume. However, there is capacitance in the system, thought to be provided by the intervertebral spaces and the vasculature. Once this capacitance is exhausted, the ICP increases dramatically with increased intracranial volume ( Fig. 24.2 ).
Based on the relation:
where CBF is cerebral blood flow, MAP is mean arterial pressure, and CVR is cerebrovascular resistance.
The concern arises that increasing ICP is necessarily associated with decrements in CBF. However, the effect of increasing ICP on CBF is not straightforward as MAP may increase with ICP elevations, and CVR adjusts with decreasing CPP (increasing cerebral blood volume) to maintain CBF until maximal vasodilatation occurs. This is thought to occur at CPP < 50 mmHg although considerable interindividual heterogeneity in this value exists and the lower limit has been nicely critiqued by Drummond. Thus, increasing ICP is often associated with cerebral vasodilatation and/or reflexly increased MAP to maintain CBF.
Normal ICP is < 10 mmHg. ICP > 20 mmHg is generally associated with escalation of ICP-reducing therapy. However, this is an epidemiologically derived number. Head trauma studies have indicated that patients with ICP > 20 mmHg generally do not do well. However, physiologically, simply elevating ICP to > 20 mmHg is not necessarily associated with decrements in CBF, provided the above-noted compensatory mechanisms occur. Recent guidelines recommend treatment of intracranial pressure when it is > 22 mmHg because values above this level are associated with increased mortality.
Nonetheless, increasing ICP because of mass lesions or obstruction of CSF outflow can exhaust compensatory mechanisms. When this occurs, compromise of CBF does eventually occur. Initially, abnormality arises in distal runoff of the cerebral circulation. As the process continues, compromise of diastolic perfusion arises. With this the normally continuous (through systole and diastole) cerebral perfusion becomes discontinuous ( Fig. 24.3 ). Further compromise of cerebral perfusion pressure results in anaerobic metabolism, exacerbation of edema, and ultimately intracranial circulatory arrest. Thus, when ICP increases it is important to detect it and ascertain whether this lethal sequence of events may be occurring.
Factors That Exacerbate Intracranial Hypertension
Under normal conditions, the intracranial volume is occupied by the brain parenchyma (80%), CSF (10%) and blood (10%). Pathologic structures, including masses caused by blood (hematoma), neoplasia, or abscesses, may also occupy space. Since the volume of the intracranial vault is somehow fixed, when the volume of one or more of these compartments enlarges sufficiently to exhaust normal capacitive compensatory mechanisms, ICP begins to rise. Further increases in volume of one or more of these compartments then leads to the sequence of events associated with increasing ICP described previously.
Two Types of Intracranial Hypertension
In a general sense, there are two types of intracranial hypertension, categorized according to CBF as hyperemic or oligemic ( Fig. 24.4 ). Although conceptualized here as a dichotomous process, undoubtedly the real physiology is more of a continuum between the two.
In the normal state, increases in CBF are not associated with increased ICP, as the normal capacitive mechanisms absorb the increased intracranial blood volume. However, in the situation of disordered intracranial compliance, small increases in intracranial volume resulting from increased CBF produce increases in ICP. When cerebral blood volume (CBV) increases, intracranial contents increase hence increasing ICP in a noncompliant system.
This, however, raises questions about an important issue. Elevated ICP has traditionally been considered to be a concern because it indicates that cerebral perfusion might be jeopardized. It is, however, unclear whether it is appropriate to be concerned about high ICP inducing intracranial oligemia when the cause of the high ICP is intracranial hyperemia. There have been no detailed examinations of this question although there have been some studies that allow reasonable inferences about the significance of hyperemic intracranial hypertension.
For many years, it has been known that brief noxious stimuli briefly increase ICP in the setting of decreased intracranial compliance. Studies have reported that such situations are associated with hyperemia, strongly suggesting that hyperemic intracranial hypertension is not a dangerous situation. However, it is reasonable to be concerned, at least theoretically, about such hyperemia for three reasons. First, elevated ICP caused by hyperemia in one portion of the brain may increase ICP to compromise CBF in other areas of the brain in which rCBF is marginal. Second, increased pressure in one area of the brain may produce gradients that might lead to a herniation syndrome. Third, there is theoretical concern that inappropriate hyperemia may predispose the brain to worsened edema or hemorrhage as occurs with other hyperperfusion syndromes. Thus, hyperemic intracranial hypertension has a theoretical potential to be deleterious but this has yet to be conclusively demonstrated. For brief periods, as may occur during intubation or other limited noxious stimuli, it is suggested that it might not be problematic.
In contrast, oligemic intracranial hypertension is associated with compromised cerebral perfusion. This is supported by the high mortality observed in head trauma patients in whom ICP rises as a result of brain edema after head injury with decrements in CBF. Transcranial Doppler (TCD) and CBF studies on these patients have demonstrated that CBF is low and perfusion is discontinuous during the cardiac cycle (see Fig. 24.3 ). Moreover, jugular venous bulb data indicates that O 2 extraction is markedly increased, suggesting anaerobic metabolism. In this setting, noxious stimuli can further increase the ICP, thus producing the situation of hyperemic or oligemic intracranial hypertension. Presumably in this setting the hyperemic rise in ICP acts to compromise rCBF further in areas of brain edema.
Although frequently linked, intracranial hypertension is not a synonym of brain herniation. These events can occur both independently or in association. Intracranial hypertension is currently defined as a sustained elevation for more than 5 minutes of ICP > 22 mmHg, and its accurate identification requires placement of an invasive intracranial monitoring. However, intracranial monitors are usually absent during the acute phase of the traumatized brain management, therefore, anesthesiologists should be familiar with clinical signs and symptoms suggestive of elevated ICP. These signs and symptoms include headache, nausea, or vomiting, altered mental status, and the Cushing triad (bradycardia, hypertension, and irregular respirations or apnea).
On the other hand, herniation syndromes are the result of differences in pressure gradients across the intracranial compartments ultimately leading to the shifting or compression of vital neural and vascular structures. Common sites of brain herniation are the medial temporal lobe (uncal), cingulum (subfalcine), and inferior cerebellum (tonsillar herniation). The classic signs of uncal herniation include loss of consciousness associated with ipsilateral pupillary dilatation and contralateral hemiparesis resultant from compression of the ascending arousal pathways, oculomotor nerve (CN III), and corticospinal tract ( Table 24.3 ).
|Syndrome||Clinical signs and symptoms||Pathophysiology|
The purpose of monitoring ICP is to improve reliably the clinician’s ability to maintain adequate CPP and cerebral oxygenation. The only way to determine CPP reliably is to monitor both ICP and MAP continuously. This combined approach has been associated with the potential to improve patients’ outcomes after closed head injury.
Despite its widespread application, supporting evidence from randomized clinical trials (RCTs) is lacking or insufficient and indications outside TBI management are somewhat ill-defined. In 2012, Chestnut et al. conducted a multicenter, controlled trial in 324 patients who suffered from severe TBI. Patients were randomly assigned to one of two specific protocols: guidelines-based management of ICP in which a parenchymal monitor was used or a treatment protocol based on clinical and imaging findings. The primary outcome was survival time, impaired consciousness, and functional status at 3 and 6 months after initial injury. Intracranial ICP monitoring failed to show superiority to care based on imaging and clinical findings. A large body of level II evidence exist with regard to ICP monitoring in TBI. Management of severe TBI patients using information from ICP monitoring is recommended to reduce in-hospital and 2-week postinjury mortality. Thus, the current recommendations are to monitor ICP in all salvageable patients with a TBI (GCS 3–8 after resuscitation) and an abnormal CT scan that reveals hematomas, contusions, swelling, herniation, or compressed basal cisterns. Also, ICP monitoring is indicated in patients with severe TBI with a normal CT scan if more than two of the following features are noted at admission: age > 40 years, unilateral or bilateral motor posturing, or SBP < 90 mmHg.
The current gold standard for ICP monitoring is the ventriculostomy catheter or external ventricular drain (EVD), where a catheter is inserted in either lateral ventricle, usually via a small right frontal burr hole. This ventricular catheter is connected to a standard pressure transducer via fluid-filled tubing ( Fig. 24.5 ). The external transducer must be maintained at a specific level to ensure adequate CPP. The reference point for ICP is the foramen of Monro, although in practical terms, the external auditory meatus is often used as a landmark ( Fig. 24.6 ). EVDs allow the clinician to measure global ICP and have some useful advantages over other ICP monitors, including the ability to perform periodic in vivo external calibration and therapeutic CSF drainage and sampling. The primary disadvantage of intraventricular catheters is the associated increased rate of infection and accidental overdrainage. This infectious risk increases the longer the EVD is in place, and prophylactic catheter changes do not appear to reduce the risk of infection. Another potential disadvantage of intraventricular systems includes a small risk of hemorrhage during placement; this risk is greater in patients taking anticoagulants or antiplatelets and misplacement ( Fig. 24.7 ). In addition, when intracranial mass lesions or ventricular effacement caused by swelling are present, EVD placement may be difficult even for the most experienced neurosurgeon.
In addition to intraventricular catheters, there is intraparenchymal, subarachnoid, and epidural ICP monitoring catheters. Intraparenchymal catheters consist of a thin wire with a fiberoptic transducer at the tip (fiberoptic Camino system, Natus Medical Inc., Middleton, WI), inserted into the brain parenchyma via a small skull burr hole. They have a lower risk of infection and hemorrhage than EVDs. Disadvantages include the inability to drain CSF for therapeutic or diagnostic purposes and the potential to lose accuracy or development of “zero drift” (degree of difference relative to zero atmosphere) over the course of days.
Noninvasive ICP monitoring devices have been developed to reduce the complications associated with invasive devices. Tympanic membrane displacement, TCD pulsatility index, intraocular pressure, optic nerve sheath diameter, and magnetic resonance imaging (MRI) of the optic nerve sheath have been used to provide a noninvasive estimate of ICP. So far, none of these methods has provided accuracy sufficient to replace invasive ICP monitors.
Normal ICP waveform consists of three arterial components superimposed on the respiratory rhythm. The first arterial wave is the percussion wave (P1), which reflects the ejection of blood from the heart transmitted through the choroid plexus in the ventricles. The second wave is the tidal wave (P2), which reflects brain compliance; and finally, the third wave is the dicrotic wave (P3), which reflects aortic valve closure. Under physiologic conditions, the percussion wave is the tallest, with the tidal and dicrotic waves having progressively smaller amplitudes. When intracranial hypertension is present, cerebral compliance is diminished. This is reflected by an increase in the peak of the tidal and dicrotic waves exceeding that of the percussion wave ( Fig. 24.8 ).
Pathologic waveforms include A, B, and C described by Lundberg in the 1960s. Pathologic A or plateau waves are abrupt, marked elevations in ICP of 50 to 100 mmHg, which usually last for 5–20 minutes and represent the loss of intracranial compliance and autoregulatory mechanisms. Moreover, they can lead to development of a positive feedback cycle that eventually leads to continuous reduction in CPP with subsequent cerebral ischemia. Thus, the presence of A waves should suggest the need for urgent intervention to help control ICP. Type B waves occur at a frequency of 0.5 to 2 waves/min with elevations of ICP of 20–30 mmHg, reflecting change in vascular tone that occurs when the CPP is at the lower limit of normal autoregulation. C waves occur at a frequency of 4–8/min; they reflect changes in systemic vasomotor tone and have little pathologic significance.
Treatment of Intracranial Hypertension
The main goal in treating intracranial hypertension is to prevent the development of secondary injuries by maintaining adequate CPP, oxygenation, and glucose supply (without hyperglycemia) to promote adequate oxygen and nutrient supplies. The best clinical strategy is to diagnose and to treat the underlying cause(s), avoid exacerbating factors, and reduce ICP ( Table 24.4 ). Current management of acute brain injury is aimed to maintain ICP < 22 mmHg, CPP between 60 and 70 mmHg, and SBP at ≥ 100 mmHg for patients 50 to 69 years old or at ≥ 110 mmHg for patients 15 to 49 or > 70 years. CPP augmentation to higher perfusing pressures may elevate the risk of systemic complications, including the development of acute respiratory distress syndrome (ARDS). More recent evidence has failed to show such association. This can be explained in part to the development of better lung-protective strategies to treat ARDS. In addition, as the Lund group has suggested in the past, hypertension-induced exacerbation of brain edema can further increase ICP. This increase in ICP leads to reduced venous outflow and increased venous pressure, which in turn act to worsen the brain edema, constituting a positive feedback cycle initially started by arterial hypertension.
|Extra-axial fluid collections|
Therapy of intracranial hypertension is primarily directed at removing the cause as far as possible. This is not always possible. In such cases, therapy is aimed at controlling ICP, hoping that the primary cause of the intracranial hypertension will resolve. Controlling ICP is thus a supportive maneuver, intended to preserve viable neuronal tissue until the high ICP situation resolves. In general, therapeutic maneuvers to reduce ICP involve one of six classes of therapy: (1) decrease cerebral blood volume, (2) decrease CSF volume, (3) induce serum hyperosmolarity, (4) resect dead or injured brain tissue or resect viable but less important brain tissue (e.g., anterior temporal lobe), (5) resect non-neural masses or hematomas, and (6) remove the calvarium to permit unopposed outward brain swelling (decompressive craniectomy). For purposes of this chapter, we will categorize interventions based on current tiered therapy algorithms ( Fig. 24.9 ). Recent advances in the use of brain tissue oxygen monitors occasionally affect the manner in which these maneuvers are used to ensure continued optimal PbrO 2 . However, their use is somewhat limited in the operating room given the invasive nature and obstruction of the surgical field.
Intracranial Hypertension/Cerebral Herniation Treatment Algorithm
Level Tier zero: L0
Head of bed elevation
Antipyretics and antishivering drugs
Airway, Breathing, and Circulation (ABCs)
The urgent assessment of airway patency, oxygenation, ventilation, and adequate circulation are particularly important in TBI to minimize development of secondary insults such as hypotension and hypoxia. If emergent intubation is required, care should be taken to minimize further elevations in ICP during intubation through careful positioning, appropriate choice of paralytic agents (if required), and adequate sedation (but not drug-induced coma). Pretreatment with lidocaine has been recommended as a useful intervention to decrease the acute elevation in ICP associated with laryngoscopy; however, good clinical evidence supporting this approach is limited.
Head of Bed Elevation
Although the beneficial effect of head of bed elevation (HBE) is unclear given the lack of consistency among existent trials, HBE should be 30 degrees when possible and kept in a neutral position to promote adequate cerebral blood venous drainage. Reverse Trendelenburg position can be utilized in the operating room if head elevation is not possible. Venous return improvement and CSF redistribution are some of the proposed benefits of HBE. The main routes for cerebral venous drainage include the deep and superficial sinus system and the internal and external jugular veins. If a cervical collar is present, ensure it is not too tight thereby limiting blood venous drainage. Providers should minimize ICP raising maneuvers (suctioning, coughing, pain). HBE should be considered when calculating CPP.
Antipyretics and Antishivering Drugs
Fever increases cerebral metabolism, thereby increasing metabolic demand and CBF, which ultimately leads to ICP elevation. Experimental studies have associated fever with increased brain injury. A recent study of 355 TBI patients evaluated the fever burden as an independent predictor for prognosis of TBI. They found that early fever might be an independent risk factor for poor prognosis in TBI.
In patients with TBI, the use of steroids is not recommended for improving outcome or reducing ICP. High-dose methylprednisolone was associated with increased mortality and is contraindicated. Moreover, glucocorticoids are not considered to be useful in the management of cerebral infarction or cerebral edema secondary to intracranial hemorrhage.
High-dose corticosteroids (dexamethasone) should be initiated if there is concern for vasogenic edema derived from brain tumors, meningoencephalitis, or abscesses. Reduction of intracranial pressure and improvement of symptoms usually occur within hours of steroid administration, and MRI changes indicating edema improvement are usually seen within 48–72 hours. Dexamethasone increases the clearance of peritumoral edema, upregulates angiopoietin-1, which is a strong stabilizing factor in the blood–brain barrier (BBB), and downregulates vascular endothelial growth factor in astrocytes. The usual initial dose of dexamethasone is 10 mg as a loading dose, followed by 4 mg every 6 hours. Patients with steroid-refractory intracranial hypertension may benefit from a carbonic anhydrase inhibitor.
In the setting of a neurologic emergency, a head CT scan should be obtained without delay to identify a process that may require immediate medical or surgical treatment. Resuscitation measures and stabilization of the patient must take place before transporting the patient to the CT scanner. Because of its rapid availability and ease to perform, CT scan is preferred over MRI in the initial setting, although MRI may be required later in the course of treatment.
Level Tier 1: L1 Cerebral Blood Volume Reduction
Hyperosmolar therapy including mannitol and hypertonic saline
External ventricular drain placement
NOTE: If L1 interventions do not successfully achieve adequate ICP control and/or there are impending signs of herniation, emergency decompressive surgery should be considered.
Hyperventilation is a quick and efficient method to lower ICP transiently in acute emergencies. In normal adults, CBV is 3–4 mL per 100 g of cerebral tissue and 70% of the total volume corresponds to venous blood. Veins and capillaries do not react to fluctuations in PaCO 2; therefore, hyperventilation-related changes in CBV are a result of arterial reactivity to changes in PaCO 2 . Hyperventilation has been utilized for almost a century and the earliest documented use was reported by Lundberg in 1959. Because CBF is largely dependent on PaCO 2 , as PaCO 2 decreases with hyperventilation, there is an associated cerebral vasoconstriction leading to a subsequent reduction in cerebral blood volume and ICP. CBF decreases by 3% for every 1 mmHg reduction in PaCO 2 and this effect is mediated by extracellular fluid (ECF) pH changes. Thus, PaCO 2 levels between 20 and 25 mmHg are associated with a reduction of CBF up to 40%–50%. However, CBF returns to its original state over 6–24 hours as the brain adapts by adjusting the bicarbonate levels in the ECF to normalize the pH. At the level of the proximal renal tubule, bicarbonate reabsorption is inhibited and excretion of H + is stimulated. These responses initiate minutes after the onset of alkalosis and persist for hours or days allowing normalization of CSF and perivascular pH. It is important to recognize that acute discontinuation of hyperventilation can cause a rebound increase in CBF leading to an ICP crisis and hyperventilation should therefore be discontinued gradually.
Hyperventilation is performed in the intubated patient by increasing tidal volume or rate, and it is commonly used to facilitate neurosurgical exposure as it provides “brain relaxation” in the surgical field. A prior multicenter randomized trial showed that hyperventilation to PaCO 2 = 25 ± 2 mmHg) was effective at reducing ICP and improving the surgical exposure during craniotomy. However, current consensus is to maintain normocapnia during intracranial surgery and to consider transient hyperventilation as a temporary measure when the “tight brain” is resistant to other means of treatment.
Cerebral and Systemic Effects of Hyperventilation
Multiple cerebral and systemic effects have been identified that can affect all organ systems, not only the cerebral vasculature. It decreases perfusion to kidneys, gastrointestinal tissue, skin, and muscle. Respiratory effects include hypocapnia-induced bronchoconstriction, reduced hypoxic pulmonary vasoconstriction, and increased permeability of the alveolo-arterial membrane. Moreover, respiratory alkalosis complicates tissue hypoxia by shifting the oxygen-dissociation curve to the left and compromises coronary blood flow, thus increasing the risk of coronary spasm. Hyperventilation has also been found to increase extracellular lactate and glutamate levels, which may contribute to secondary brain injury.
Hyperventilation introduces a risk of decreasing CBF to a dangerous level ( Fig. 24.10 ). A study in trauma patients showed that routine hyperventilation was associated with worse neurologic outcome at 3 and 6 months after the injury ( Fig. 24.11 ). The reason for this is uncertain as no-one has demonstrated that anaerobic metabolism occurs with hyperventilation in this setting. Possible causes are: (a) HV produces alkalemia and increased affinity of oxygen for hemoglobin, (b) it decreases seizure threshold, (c) the potential exists for hyperventilation to produce only a transient effect or to produce a paradoxical increase in ICP ( Fig. 24.12 ), and (d) upon discontinuation of hyperventilation, a paradoxical CSF acidosis can occur.
Multiple clinical studies in TBI have confirmed that hyperventilation can cause significant reductions in CBF. Jugular bulb venous oxygen saturation (SjvO 2 ) and partial brain tissue oxygen pressure (PbtO 2 ) values can be dramatically reduced by continuous hyperventilation. Hypocapnia increases the likelihood of detecting brain tissue hypoxia when using PbtO 2 monitors. This effect is more pronounced during the first few days post-TBI and it is associated with poor outcomes. Using a modified PET scan, Coles et al. evaluated CBF, CBV and oxygen extraction factor (OEF) in patients within 10 days of trauma. Hypocapnia was found to cause a decrease in CBF, and an increase in volume of ischemic areas and OEF.
The presence of hyperemia can be determined by the use of direct brain CBF determination or via jugular bulb oximetry. Brain tissue PbrO 2 may be helpful in this determination but this remains to be definitively demonstrated. High ICP associated with a low A-VDO 2 across the brain (3–4 vol%) is thought to indicate that hyperventilation can be safely used. In an emergency situation, even if the nature of the high ICP (oligemic vs hyperemic) is not known, acutely administered hyperventilation should be used to keep ICP down or to reverse a herniation syndrome or plateau wave until more definitive diagnosis or therapy can be performed.
Hyperventilation in TBI
A post-traumatic brain is extremely sensitive to ischemic damage. The autoregulatory pressure and the CO 2 reactivity mechanisms are usually exacerbated during the acute phase of TBI, especially in penumbral areas adjacent to cerebral contusions or hematomas. Loss of pressure autoregulation and CO 2 reactivity is usually associated with poor TBI outcomes. For these reasons, it is vital to maintain CPP while avoiding hypotension and hypocapnia-induced vasoconstriction.
No recent studies have established the direct association of hyperventilation and clinical outcome after TBI. In 1971, Gordon described a large retrospective series of 251 patients treated with prolonged hyperventilation, 51 of whom were hyperventilated to a PaCO 2 between 25 and 30 mmHg for a time period that varied between 6 hours and 41 days (mean 10 days). The hyperventilation group had a lower mortality (9.8% vs 32.8%); however, survivors had increased severe neurological sequelae. Among patients who experienced a complete recovery, there was no difference between groups. Another prior prospective, randomized study evaluated TBI outcome in patients who were treated with and without hyperventilation for 5 days. Favorable outcomes toward the hyperventilation group were seen at 3 and 6 months; however, after a year the differences were no longer significant.
Target PaCO 2 Recommendations in TBI
Despite the fact that there is insufficient evidence to establish clearly whether hyperventilation is beneficial or deleterious after TBI, current guidelines recommend against its prophylactic use.
In emergency situations, a short period of 10–15 minutes of hyperventilation to a PaCO 2 of 30–35 mmHg is recommended to treat acute intracranial hypertension.
Targeted PaCO 2 during normoventilation is 35–40 mmHg with a pulse oxygen saturation of ≥ 95% and/or PaO 2 of ≥ 80 mmHg.
Hyperventilation should be avoided during the first 24 hours after injury when CBF often is reduced critically and the risk of ischemia is higher.
If hyperventilation is used, SjvO 2 or PbtO 2 measurements are recommended to monitor oxygen delivery.
Do not stop hyperventilation suddenly because it may cause rebound ICP elevation.
If CBF is known and observed to be hyperemic, contributing to intracranial hypertension, there is a theoretical basis for hyperventilation used in this personalized manner, but with continued CBF monitoring. Such monitoring is not currently widely available but may be an element of future routine care.
Hyperventilation is recommended as a temporary measure to reduce high levels of ICP in the following situations :
Herniation syndromes as described previously
Life-threatening elevations of ICP. For example, type A plateau waves, while investigating triggers and expecting the effect of osmotherapy.
Refractory intracerebral hemorrhage (ICH). Hyperventilation is used in conjunction with other tier-3 level measures, such as decompressive craniectomy, hypothermia, or high doses of barbiturates.
To facilitate brain relaxation during neurosurgical interventions only when other measures have failed.
The intravascular administration of hyperosmolar agents such as mannitol and/or hypertonic saline (HTS) creates an osmolar gradient that facilitates water transport across the BBB into the circulation, where it is excreted by the kidneys. The net effect is a reduction of the interstitial fluid and a decrease in ICP. For acute elevation in ICP, therapy with either mannitol or HTS have shown similar efficacy in decreasing ICP.
Mannitol remains a commonly used hyperosmolar agent. It is an osmotic diuretic freely filtered by the renal glomerulus and does not undergo tubular reabsorption. It lowers ICP through two main effects; first, through hypoviscosity-mediated autoregulatory vasoconstriction, also called the “rheological effect,” and second, it may induce a further decrease in ICP through brain dehydration in areas with intact blood brain barrier. However, this may be limited through generation of intracellular idiogenic osmoles equalizing transmembrane osmolar gradients.
Mannitol is administered as 0.5–1 g/kg as intravenous (IV) bolus through either peripheral or central line and can be repeated every 4–6 hours. Reduction in ICP is usually achieved within minutes, peaks at about 1 hour, and last 4 to 6 hours. Fluid balance, renal function, and plasma osmolality must be monitored because no therapeutic effect is seen with osmolality > 320 mOsm/kg. If plasma osmolality is > 320 mOsm/kg, osmolar gap (measured osm – calculated osm) should be calculated and if < 20 mOsm/kg, mannitol should be administered. If > 20 mOsm/kg, HTS should be considered, because the risk of renal dysfunction is higher. Patients with known renal disease may be poor candidates for osmotic diuresis. Another frequent complication related to mannitol is hypotension because of its diuretic effect. Acute hypotension is most commonly seen after rapid infusions and can be minimized by prolonged infusions (15–30 minutes).
Unfortunately, mannitol can have delayed effects to increase ICP. This can occur by four mechanisms. First, as a potent osmotic diuretic, mannitol can have a secondary effect to decrease systemic blood volume, thus decreasing cardiac output and blood pressure. This can result in normal reflex autoregulatory increases in CBV, which can increase ICP. Second, the increased urine output, if not replaced with commensurate IV fluid therapy, can elevate hematocrit, thus opposing the initial mannitol-induced decrease in viscosity. Third, mannitol can cross the BBB in an unpredictable manner with the possibility introduced of rebound increase in ICP, similar to that observed with urea. This is partly related to the reflection coefficient of 0.9 indicating that it can even slowly diffuse into the normal brain. Fourth, there is a theoretical possibility of generation of increased intracellular osmolarity, via so-called idiogenic osmoles, which may predispose to rebound increase in brain volume with discontinuation of mannitol. These complications of mannitol are probably lessened if urine output is replaced with balanced crystalloid infusion and if, once blood osmolarity is increased, it is not allowed to decrease rapidly to the prior level unless clinical improvement indicates that weaning of ICP-reducing therapy is appropriate. This certainly should be considered in any patient demonstrating periodic abrupt increases in ICP.
HTS has increasingly been used for treatment of acute elevated ICP, replacing mannitol as a first-line agent in multiple institutions. With a reflection coefficient of 1.0 and no potential to produce deleterious diuresis and undesired hypovolemia, it has properties that made it the most attractive hyperosmolar agent.
The ability of HTS to lower ICP was initially described in 1919 by McKesson and it is available in different concentrations ranging from 2% to 23.4%. HTS can be given as a bolus alone or in addition to mannitol. HTS 2% and 3% boluses can be given via a peripheral line and concentrations > 5% should be given via a central venous catheter because of elevated osmolarity ( Table 24.5 ). However, in acute emergencies 3%–7.5% HTS should be administered regardless of the presence of central venous access. Two prior prehospital trials using HTS 3% and 7.5% via peripheral lines did not show any negative effects. Administration of HTS through intraosseous access should be done carefully and with concentrations of < 7% because of the potential risk of myonecrosis.
|Mannitol 20%||N/A||1098||0.5–1 g/kg|
|Mannitol 25%||N/A||1375||0.5–1 g/kg|
|HTS 3%||513||513||1027||150 mL|
|HTS 5%||856||856||1711||150 mL|
|HTS 23.4%||4004||8008||30 mL|
|Normal saline 0.9%||154||154||308||N/A|
The mechanism of action behind HTS involves many theories, with the most common involving the creation of an osmotic shift of fluid from the intracellular space to the intravascular and interstitial space. In addition, HTS shrinks red blood cells, making them more deformable and enhancing their passage across capillaries. A recent study shows that HTS can reverse brain oxygenation and metabolism dysfunction through vasodilatory, mitochondrial, and anti-edema effects. The combination of these effects result in a biphasic reduction in ICP by improving rheologic properties and then by osmotic activity through aquaporins across the BBB. Aquaporins are water channels present in the brain responsible for CSF production and water distribution; its downregulation after TBI is associated with development of cerebral edema. In addition, prior evidence suggests that HTS modulates the innate immune response by diminishing the neutrophil and endothelial cell activation leading to a reduction in microvascular injury and permeability.
HTS has been studied in animal models of ICH, subarachnoid hemorrhage (SAH), and TBI. In a notable prior related canine ICH study, HTS was associated with decreased intraparenchymal pressure throughout the brain, including the perihematomal region. More recently, Schreibman et al. described an experimental ICH model in which hyperosmolar therapy shows reduction of neuroinflammation by causing a reduced activation of microglia/macrophages in perihematomal tissues. Yin et al. suggest that HTS ameliorates TBI-induced cerebral edema by suppressing brain edema, proinflammatory cytokine expression and apoptosis via downregulation of aquaporin 4 in a rat TBI model. In dogs with ICH treated with either mannitol or HTS, both drugs effectively decreased ICP although HTS duration of action was longer.
In humans, the effect of HTS has been reported in ischemic stroke, ICH, SAH, TBI, and hepatic encephalopathy. All studies show that HTS effectively and reproducibly reduces ICP with concomitant improvement in CPP. Indeed, Suarez et al. showed one important effect of 23.4% HTS in eight patients as a therapy that effectively decreases ICP when all other medical therapies do not work ( Fig. 24.13 ). Tseng et al. reported that in 10 patients with poor grade SAH, ICP was found to decrease by around 75% and CBF was enhanced in the face of decreased cerebrovascular resistance, indicating potential rheologic and cerebral vasodilatory of HTS 23.5%. Al-Rawi et al. demonstrated that an increase in CBF is followed by improvement in tissue oxygenation and metabolism when poor-grade SAH patients are treated with 23.5%. More importantly, this increment in tissue oxygenation was found to correlate with improved neurologic outcomes.
Ware et al. assessed the effects of 23.4% HTS in TBI patients with elevated ICP refractory to mannitol. They found HTS to have a longer effect on ICP reduction than mannitol and to be effective in patients with Na > 150 mEq/L. Paredes-Andrade et al. studied HTS 23.4% on ICP in the setting of different serum and CSF osmolarities. HTS was found to decrease ICP irrespective of serum or CSF osmolarity, possibly indicating a nonosmotic mechanism of action. More recently, a 30% HTS concentration was tested in TBI and no associated harmful or hematologic abnormalities were observed. In 2013, Eskandari et al. tested HTS 14.6% after TBI and showed that patients who exhibited refractory intracranial hypertension can be treated effectively and safely with repeated boluses of 14.6% with no significant change in renal function or hemodynamics. Oxidative stress processes play an important role in secondary brain injury. In a study of 33 adults with TBI, HTS showed antioxidant effects by decreasing the amount of serum total antioxidant power, reactive oxygen species and nitric oxide. Diringer et al. studied the effects of HTS 23.4% and mannitol 20% on CBF, CBV, OEF, and cerebral metabolic rate (CMR) O 2 in nine patients who developed cerebral edema with midline shift from acute ischemic stroke. They found that increased CBF on the contralateral unaffected hemisphere was dependent on mean arterial pressure. No effect on CBV was observed, arguing against the compensatory cerebral vasoconstriction previously proposed as a mechanism for ICP reduction.
Efficacy of HTS for Reduction of ICP
Several small studies in humans have assessed the ability of HTS to reduce ICP. A retrospective study by Qureshi et al. showed that HTS decreases ICP in head trauma in postoperative brain edema but not nontraumatic ICH or ischemic stroke. Schatzman et al. performed a prospective nonrandomized evaluation of HTS in severe head injury and found that it effectively reduced ICP. A subsequent prospective randomized study by Vialet et al. in TBI patients found better ICP control compared with mannitol. Another prospective randomized study did not show better ICP control with HTS compared with standard therapy. However, in this study the sample size was relatively small and the HTS patients were sicker on entry into the study. Three studies report that HTS can be safely and effectively used in children to decrease ICP after TBI. Suarez et al. studying severe SAH patients reported that HTS therapy effectively decreased ICP while concomitantly increasing CBF (see Fig. 24.13 ). Schwartz et al. evaluated HTS/hetastarch therapy in ischemic stroke patients with high ICP, compared with mannitol. Both therapies decreased ICP, but the HTS group had better control.
Patients with intracranial hypertension associated with hepatic encephalopathy have also been demonstrated by Murphy et al. to sustain ICP decrements after HTS. Moreover, HTS and mannitol have been reported as a useful bridge therapy while awaiting liver transplant in patients who suffer from cerebral edema.
Although many studies show the efficacy of HTS in improving ICP and other parameters, the effect of this intervention on long-term clinical outcomes remains unclear. In 1991, Vassar et al. evaluated this hypothesis showing a trend for better outcomes in HTS-treated patients. One controlled trial randomly assigned 226 patients with TBI to prehospital resuscitation with 250 mL HTS 7.5% or the same volume of Ringer lactate. Survival until hospital discharge, 6-month survival, and neurologic function 6 months after injury were not different. This study did not assess the impact of systematic use of HTS throughout the hospital course. Koenig et al. evaluated the effect of 23.4% HTS on reversal of transtentorial herniation in 68 patients who suffered from various intracranial pathologies. HTS reversed transtentorial herniation quickly and reduced ICP, but clinical outcomes remained poor.
Overall, mannitol and HTS have been compared in some randomized trials of patients with intracranial hypertension derived from a variety of causes (traumatic brain injury, stroke, tumors). Meta-analyses of these trials have found that HTS appears to have greater efficacy in managing elevated ICP, but clinical outcomes have not been systematically examined. Further clinical trials are required to clarify the appropriate role of hypertonic saline infusion versus mannitol in the acute management of intracranial hypertension.
HTS clearly has the potential to exert a positive impact in the management of intracranial hypertension. However, there are concerns expressed about possible deleterious effects. Perhaps one of the most worrisome concerns is renal failure. This was mentioned in a report by Peterson wherein 2 of 10 children developed reversible renal failure, also, however, related to multiorgan failure and sepsis. Nonetheless, the issues were further underscored in editorials by Valadka et al. and Dominquez et al. but with disagreement about this as a significant risk expressed by Bratton et al. Other potential adverse effects were reviewed by Suarez and include:
Disruption of the BBB (“osmotic opening”).
Excess neuronal death. Postulated after continuous infusion of 7.5% saline in a rat model worsened outcome after transient ischemia. This has not been found to be clinically relevant.
Alterations in the level of consciousness associated with hypernatremia.
Central pontine myelinolysis. This is typically associated with too-rapid correction of (chronic usually) hyponatremia. It has not been associated with the use of HTS in humans from a normal sodium concentration.
Congestive heart failure.
Transient hypotension. This has been reported inanimals after rapid intravenous hypertonic fluid infusions.
Decreased platelet aggregation and coagulation factorabnormalities of HTS. However more recent evidence failed to show this effect in coagulation using rotational thromboelastometry (ROTEM) to assess coagulation and platelet function. Patients receiving HTS or mannitol for ICP control did not exhibit impaired coagulation function.
Hypokalemia and hyperchloremic metabolic acidosiswith large quantities of HTS.
Phlebitis. Infusion should be done via a central venouscatheter.
Renal failure. As discussed previously, the relationship to HTS was not clear in the report in which this was described. There are no reports of renal failure in animals or humans in the absence of more common ICU etiologies of renal failure, such as sepsis and organ failure.
CSF volume is reduced by removal via ventricular drain (see Figs. 24.5 and 24.6 ). This can be affected by setting the drainage at a prescribed level above the midbrain or prescribing that the drain be opened anytime ICP exceeds 20–25 mmHg. Leaving a drain open risks excessive CSF drainage when the patient coughs or with drain manipulation in the course of routine nursing procedures and can contribute to collapse of the ventricles or subdural hemorrhage. Excessive and abrupt decrease in local pressure around the drain can produce intracranial gradients, leading to a herniation syndrome. Leaving the drain clamped and monitored, however, risks the development of untreated intracranial hypertension.
Level Tier 2: L2
Hypertonic saline infusion: If HTS has been initiated,sodium goal should be established. In general, a sodium level > 160 mEq/L is unlikely to provide any extra benefit. In order for HTS to be an effective ICP control measure, an intact BBB and a sodium gradient (Na serum -Na brain ) must be present to promote the escape of water from the brain tissue. If adequate ICP control is achieved with HTS infusion, the serum sodium concentration at which ICP was controlled should be maintained until cerebral edema has subsided. (See Tier 1 section on hyperosmolar therapy for detailed information.)
Cerebral blood volume reducing drugs: propofol, etomidate, lidocaine.
CBV Decreasing Drugs
CBF decreasing (and therefore CBV decreasing) drugs that can decrease ICP include barbiturates, benzodiazepines, etomidate, and propofol. Notably, all are CNS depressants and become ineffective once the EEG become isoelectric. Thus, their use indicates acceptance on the part of the clinician to lose any reliable neurologic examination. Unlike hyperventilation, these agents decrease CBF coupled to CMR. Thus, the CBF decreases should not provide a milieu for anaerobic metabolism. Lidocaine also decreases CBF and CMR to decrease ICP although with a less pronounced decrement in neurologic function. Mannitol’s immediate effects are also thought to be mediated by reduction in cerebral blood volume, although this is undoubtedly minor and temporary as a mechanism. Barbiturates are discussed as a tier-3 therapy in the next section.
Propofol has been shown to reduce ICP by decreasing CMRO 2 and CBF volume. In acute situations, it can be given as a bolus of 1–3 mg/kg and continued as an infusion in mechanically ventilated patients. Although propofol has some ICP-reducing properties it also decreases blood pressure significantly, especially when given as a bolus, which should be corrected with fluids or vasopressors aimed at maintaining CPP. Its use should be guided by bispectral index (BIS) monitoring or other forms of EEG assessment. Recent evidence shows that BIS is more reliable than Richmond Agitation and Sedation Scale (RASS) for maintaining a stable sedation status and ICP. Deeper sedation levels (BIS 40–50) are associated with a faster decrease in ICP. A small subset of patients treated with propofol may develop a propofol infusion syndrome (PRIS). PRIS is a collection of clinical and laboratory findings that develop after propofol infusion and are characterized by metabolic acidosis, elevated CK, hyperkalemia, elevated liver enzymes, rhabdomyolysis, triglyceride elevation, and acute renal failure. PRIS is more frequently seen after long-term (> 48 hours) or high-dose infusion > 4 mg/kg/h or 67 μg/kg/min. Management includes immediate discontinuation of propofol. Its use should be limited to no more than 48 hours. Cardiac dysfunction and arrhythmias are a major cause of mortality. Cardiogenic shock should be managed with inotropic support and mechanical devices and extracorporeal support in severe cases. If propofol is infused at these extreme rates (200 μg/kg/min) it should only be done temporarily while other corrective measures are executed. Children suffering severe TBI appear to be particularly sensitive to PRIS. Thus, this ICP lowering strategy is frequently avoided.
Etomidate decreases ICP by decreasing CBF and CMR. Its hemodynamic effects are not as potent as those of barbiturates with the result that it can decrease ICP without decreasing CPP as much as with thiopental. It is not suitable for prolonged infusion to control ICP because of inhibition of adrenal corticosteroid synthesis unless steroids are concomitantly administered and because of the possibility of renal compromise resulting from propylene glycol toxicity. Nonetheless it can be used for brief periods as an adjunct in ICP control during anesthesia induction, especially if there is hemodynamic concern. The dose is 0.1–0.3 mg/kg IV.
Lidocaine decreases ICP by decreasing CBF and CMR but without as much CNS depression as barbiturates. It is not as potent or reliable as barbiturates in decreasing ICP but it can be useful to decrease ICP when there is hemodynamic instability or when barbiturates are thought to be not tolerated hemodynamically. It can be useful prior to airway manipulations but is not typically used in prolonged therapy of intracranial hypertension. It is most useful when given 0.5–1.5 mg/kg IV or intratracheally for acute treatment of high ICP, particularly if associated with airway manipulation.
Level Tier 3: L3
Tier 3 measures account for the most aggressive level of treatment in acute intracranial hypertension and as such carry the most serious side effects. Interventions include:
transient hyperventilation to PaCO 2 25–35 mmHg.
The use of barbiturates to decrease ICP is based on their ability to reduce CBF, CBV, CMR, and seizure activity. Pentobarbital infusion is generally used with a loading dose of 5 to 20 mg/kg as a bolus, followed by maintenance infusion of 1 to 4 mg/kg/h titrated to ICP/CPP goal or burst suppression of 5–20 seconds on continuous electroencephalogram (EEG). EEG burst suppression is an indication of maximal dosing. Increasing the barbiturate dose beyond that needed to produce burst suppression is unlikely to provide further decreases in ICP because further decrements in CMR are not thought to occur at doses beyond that needed for significant burst suppression. Alternatively, serum pentobarbital levels can be monitored aiming for 30–50 μg/mL. The pentobarbital infusion is continued for 24–96 hours while the processes driving elevated ICP are treated. Discontinuation of barbiturate therapy can be considered when (a) there is no decrease in ICP with barbiturate loading, (b) despite an initial favorable response, intracranial hypertension recurs despite maximal barbiturate therapy, or (c) ICP has remained below 15 mmHg for longer than 24–48 hours. In the latter case barbiturates should be weaned with continued ICP monitoring.
The therapeutic value of this “barbiturate coma” is somewhat unclear. In a prior randomized trial of 73 patients with intracranial hypertension refractory to standard therapy, patients treated with pentobarbital were 50% more likely to have their ICP controlled. However, there was no difference in clinical outcomes between groups. Two other RCTs reported that barbiturates did not significantly affect mortality outcome in TBI. Ward et al. assigned 53 patients to either prophylactic pentobarbital treatment versus no pentobarbital treatment; there was no survival difference. Moreover, higher incidence of hypotension was observed in the pentobarbital group. Schwartz et al. studied the effect of mannitol and pentobarbital on ICP control on patients with and without hematoma after TBI. In both cohorts ICP control was worse with pentobarbital. In general, the use of barbiturates is a “last-ditch” effort, because several studies show that their ability to lower ICP does not appear to affect outcomes. The most recent Brain Trauma Foundation Guidelines recommend against administration of barbiturates to induce burst suppression measured by EEG as prophylaxis against the development of intracranial hypertension. Additionally, high-dose barbiturate administration is recommended to control elevated ICP refractory to maximum medical and surgical treatment. Hemodynamic stability is essential before initiation of therapy.
There are several adverse effects of prolonged barbiturate therapy for high ICP. Barbiturates blunt the neurologic examination. Thus, a growing intracranial mass lesion or other neurologic exacerbation may not be obvious until it causes an increase in ICP or increasing barbiturate requirement to maintain normal ICP. Hypotension and respiratory depression may occur. Blood pressure may need to be pharmacologically supported and intubation and mechanical ventilation are mandatory. Dysfunction of the alimentary tract may occur making it difficult to use enteral nutrition. Thermoregulation is disturbed. Hypothermia may occur inadvertently and infection may be masked because there is no fever. Finally, physical addiction and tolerance may occur with prolonged use making it difficult to discontinue barbiturate treatment without seizure.
Thiopental 1–4 mg/kg can be given, repeated as needed, to control ICP. It is subsequently infused as needed to control ICP. In this setting, thiopental has been reported to produce hypokalemia with induction and rebound hyperkalemia on drug cessation
See section on temperature control.
Selected patients with evidence of rapid neurological deterioration from focal-space occupying lesions and those who failed medical management may benefit from surgical decompression, even if medical therapy has not been attempted. The following are procedures considered for different emergent scenarios.
Evacuation of extra-axial collection (i.e., subdural or epidural hematoma).
Resection of intracerebral lesions caused by abscesses or lobar hemorrhages. Occasionally, lobectomies can be performed as life-saving procedures.
Posterior fossa decompression in the event of obstructive hydrocephalus resultant from vasogenic edema caused by tumors and ischemic or hemorrhagic strokes.
Uni- or bilateral decompressive hemicraniectomy after traumatic brain injury or malignant ischemic stroke.
Note: Surgical decompression can be performed at any point in the tiered algorithm at the discretion of the neurointensivist and neurosurgeon, if conventional therapy fails to improve ICP or evidence of impending herniation exists.
First described by Thomas Kocher in 1901 and subsequently by Cushing in 1908, decompressive craniectomy (DC) is a surgical procedure in which part of the skull uni- or bilaterally is removed and the underlying dura opened. DC removes the rigid confines of the cranial vault, allowing further expansion of the intracranial contents with subsequent reduction of intracranial hypertension and brain herniation risk. Although intravascular arterial blood accounts for the majority of the intracranial blood volume, venous contribution to ICP is commonly overlooked. Diffuse brain edema can lead to generalized venous compression and a cycle of venous hypertension. In a recent study after TBI, alleviation of venous sinus compression might have contributed to the ICP-reducing effect of DC.
Although DC has been investigated in a number of neurological conditions, RCTs have only been evaluated in TBI and stroke. DC can be a primary or secondary procedure. A cohort study of patients with traumatic subdural hematoma (SDH), demonstrated lower mortality in patients undergoing primary DC compared with conventional craniotomy where the bone flap is replaced. However, there is no high-quality evidence to support this approach. The Randomized Evaluation of Surgery with Craniectomy for patients Undergoing Evacuation of Acute Subdural Haematoma (RESCUE-ASDH) trial is a multicenter, randomized trial comparing the effectiveness of primary DC versus craniotomy and bone-flap replacement after evacuation of an ASDH in adult head-injured patients ( www.rescueasdh.org/ ); subjects recruitment was completed in 2019. Results of trial are still pending. Secondary DC is most commonly undertaken as a last-tier (life-saving) intervention in a patient with severe intracranial hypertension refractory to tiered escalation of interventions to control ICP. There are three main approaches to DC (bifrontal, uni- or bilateral hemicraniectomy, suboccipital). The craniectomy must be of sufficient size to allow effective ICP control.
Decompressive Craniectomy and TBI
Nonrandomized trials have shown that DC is an effective therapy to control intracranial hypertension and to improve survival after severe TBI. However, it seems to be at the expense of long-term functional outcome with variable degrees of severe disability. Two recent randomized clinical studies—the Decompressive Craniectomy (DECRA) study and the Randomized Evaluation of Surgery with Craniectomy for Uncontrollable Elevation of Intracranial Pressure (RESCUEicp) study —investigated DC to control elevated ICP after severe TBI. The DECRA study randomized 155 adults to either bifrontal craniectomy or standard care for elevated ICP. DC was associated with lower ICP than medical treatment, but the long-term outcome was unfavorable (vegetative state or severe disability). The RESCUEicp study investigated the effect of bifrontal craniectomy or unilateral hemicraniectomy as a last-tier therapy for refractory intracranial hypertension. A study was made of 408 subjects randomized to either continuing medical treatment or DC if ICP > 25 mmHg for at least 1hour refractory to medical treatment. Compared with medical management, DC resulted in lower mortality at 6 months (48.9% vs 26.9%, respectively; P < 0.01), but higher rates of vegetative state and severe disability.
Decompressive Craniectomy and Stroke
Several trials have investigated the effect of DC on malignant MCA stroke. The Decompressive Craniectomy in Malignant MCA infarction (DECIMAL), the Decompressive Surgery for the Treatment of Malignant Infarction of the Middle Cerebral Artery (DESTINY) trial, and Hemicraniectomy After Middle Cerebral Artery infarction with Life-threatening Edema Trial (HAMLET). The three trials confirmed the mortality benefit of DC in patients with malignant MCA stroke < 60 years old, but with no individual improvement in functional outcome. Hemicraniectomy in patients > 60 years old was investigated in the DESTINY II trial. Compared with medical treatment, hemicraniectomy within 48 hours of malignant hemispheric infarction resulted in lower mortality (33% vs 70%; P < 0.001) but a higher proportion of severely disabled survivors (mRS 4–5) in patients between 61 and 82 years of age. Two systematic reviews have summarized clinical findings after DC in stroke; irrespective of age, DC significantly reduces mortality and improves functional outcome in adults with malignant MCA infarction but with a nonsignificant increase in the risk of survival with major significant disability.
DC is a major surgery and it comes with associated potential complications including subdural hygromas, infection, hydrocephalus, cerebral herniation through skull defect, CSF leak, and ipsi- or contralateral EDH/SDH. The “syndrome of the trephined” initially described by Grant and Norcross in 1939 has become more frequently associated with the resurgence of DC for intracranial hypertension. It often occurs after a period of recovery and consists of seizures, weakness or paralysis, sensory changes, altered mental status, and headache. The treatment is cranioplasty.
A much less common but potentially lethal complication is paradoxical transtentorial herniation after hemicraniectomy with a large skull defect. This may occur at any time after DC, even after several months. DC renders the supratentorial space vulnerable to atmospheric pressure, which must be buffered across potential herniation sites by CSF pressure (i.e., foramen magnum or skull defect). When this buffer is compromised by CSF leaks, lumbar puncture, or ventriculostomy drains, herniation can occur. Usual symptoms include those of acute ICP elevation such as decreased level of consciousness, anisocoria, bilateral pupillary dilatation or loss of reactivity, autonomic instability. However, treatment aimed toward ICP control will not be effective and may aggravate the situation. Instead, the patient should be placed supine or in the Trendelenburg position, CSF drains should be clamped, crystalloid fluid should be administered intravenously, and an epidural blood patch should be placed for patients with dural CSF leak.
Decompressive craniectomy has a major role in treating intractable intracranial hypertension in severe TBI and malignant stroke, but its meaningful benefits remain unclear. Although DC is considered an option for ICP control, it should not be routine for management of intracranial hypertension. Decision to perform DC should encompass not only the acute scenario, but also the most likely long-term outcome. The patient’s family should be informed extensively about this procedure before a decision is taken.
The essence of satisfactory perioperative neurologic care is to provide a physiologic and biochemical milieu that will promote a good recovery.
Central to physiologic considerations in management of the acutely injured brain are issues related to cerebrovascular reserve. Simply stated, this refers to the capacity for the brain to compensate successfully for physiologic stresses such as hypoglycemia, hypoxemia, hypotension, and anemia. In all of these situations, vasodilation arises to provide a compensatory increase in CBF.
Animal experiments indicate that it is possible to produce a condition in which cerebrovascular reserve is compromised with increased tendency to cerebral infarction. For example, occlusion of one carotid artery or production of moderate hypoxemia do not produce symptoms on the own because cerebral vasodilatation occurs to compensate. Indeed, some contend that arterial hypoxemia, occurring with normal cerebral vascular compensatory mechanisms, does not cause brain damage. Of course, one contributing factor to this notion is that hypoxic myocardial dysfunction produces organismic death such that isolated neuronal injury cannot occur. However, add hypoxemia to carotid occlusion, or vice versa, and a stroke occurs because compensatory mechanisms, already fully utilized, cannot accommodate the further decrease in O 2 supply. Examples of variants of this situation abound clinically. Such examples of attenuated cerebrovascular reserve include cerebral edema, hypoxemia, anemia, carotid stenosis, peri-infarct penumbra, and so on. In each of these situations, although not easy to quantitate, it is clear that added compromise of O 2 supply to the brain risks neuronal injury.
Airway Evaluation and Management
In patients with a damaged central nervous system, there are several concerns with respect to airway management. First, if there is intracranial pathophysiology, the associated hemodynamic or respiratory changes associated with airway manipulation can lead to elevated ICP, exacerbated brain edema, or worsened intracranial hemorrhage. Although many patients arrive in the operating room already intubated, some, particularly those with extradural hematoma, may be conscious and breathing spontaneously. Airway management in TBI can be complicated by several factors, including urgency of situation, uncertainty of cervical spine status, uncertainty of airway obstruction (caused by presence of blood, vomitus, debris in the oral cavity, laryngopharyngeal injury, or skull base fracture), full stomach, intracranial hypertension, and uncertain volume status. All patients with TBI requiring urgent surgery must be assumed to have a full stomach. If there is spinal cord injury, which may not be initially clinically apparent, then cervical spine manipulation risks producing or worsening any acute plegic deficits. Most head-injured patients are assumed to have an unstable cervical spine until definitive clearance can be obtained.
Evaluation of the airway is performed as for most preanesthetic evaluations. However, specific items need to be sought in the history and physical examination:
NPO status. This is helpful history but impacts little on plans as most acutely injured patients are assumed to have a full stomach.
Evidence of elevated intracranial pressure. This may affect choice of drugs used to support intubation. (See previously for clinical signs and symptoms of intracranial hypertension.)
History of the incident. This can lead to approaches based on the cause of the injury being direct trauma or perhaps a medical problem that led to trauma.
Other medical history if available.
Presence of neurologic deficits, especially paraplegia or quadriplegia. This may impact on the choice of drugs to support intubation and the methods by which intubation is performed, or it may suggest possible autonomic dysfunction. Quadriplegic patients have a significant attenuation of the catecholamine response to intubation.
Evidence of airway injury.
Evidence of aspiration or other impediments to adequate gas exchange. Associated lung injuries or aspiration related to altered mental status will lead to faster desaturation during intubation and probably worsen the neurologic outcome.
The main options to consider in airway management have to do with whether and how deeply to produce CNS depression and which tool will be used to effect intubation of the trachea. There are pros and cons to each approach and considerable clinical judgment is needed to match the spectrum of benefits and risks to the underlying disease processes.
Awake Versus Asleep
There is a continuum of anesthetic states that may be used, varying from awake and unsedated to fully anesthetized with neuromuscular blockade. In the context of possible intracranial hypertension or hemorrhage, the side effects of anesthetic drugs can be usefully used to control systemic and intracranial hypertension. The full spectrum of effects of anesthetic drugs is beyond the scope of this chapter. But, briefly, thiopental, propofol, and etomidate have direct ICP-reducing qualities that can be of use during intubation, a procedure well-known to produce intracranial hypertension. Attenuation of systemic hypertension can be augmented with judicious use of an opioid such as fentanyl in addition to a blood pressure–reducing drug such as thiopental or propofol. Neuromuscular blockade can be achieved with vecuronium or rocuronium, neither of which has any reported cerebrovascular effects. Succinylcholine can also be used although concerns remain about it causing increased ICP, probably related to increased cerebral blood flow inducing afferent spindle stimulation. However, after a large bolus of one of the usual aforementioned induction drugs, succinylcholine seems to be associated with no significant problems. Moreover, its use is theoretically safe with hyperacute onset paraplegia or quadriplegia (i.e., before nicotinic receptor upregulation) but Dr. Cruz and Dr. Kofke refrain from its use at any time with spinal cord injury patients because of concern about the speed of receptor upregulation or inaccurate history leading to fatal hyperkalemia.
Awake endotracheal intubation is an option for cooperative patients with acute spinal cord injury (SCI). However, its use has advantages and disadvantages ( Table 24.6 ). Retrospective studies have not shown differences in neurologic outcome with awake versus asleep fiberoptic intubation in acute SCI. Thus, the choice of technique should be based on clinical situation, patient cooperation, and clinical expertise.
Cricoid or No Cricoid Pressure?
It has been hypothesized that any unidirectional force applied to the cervical vertebrae during cricoid pressure may cause excessive neck movement and exacerbate preexisting lesions. Studies have demonstrated that cricoid pressure causes cervical spine displacement of different degrees when the posterior aspect of the neck is not supported. The clinical implication of these movements was assessed retrospectively in patients who had C-spine injuries and they were found to be free of neurologic impairment. However, it would be sensible to avoid movements of the potentially fractured cervical spine, if at all possible. The double-handed maneuver, which is popular in trauma patients to provide support to the posterior cervical spine and to provide stabilization to the posterior aspect of the neck, seems to be a safer alternative to the single-handed maneuver. The assistant performing bimanual CP should not be assigned to other duties until intubation has been completed. If the bimanual technique is obstructing the glottic view, it may need to be switched back to a single-handed maneuver to improve laryngoscopic visualization. Cricoid pressure utilization is controversial in SCI because forceful pressure over the site of the C-spine fracture can cause displacement. Moreover, the efficacy of cricoid pressure to prevent regurgitation has been questioned. These authors do not routinely use cricoid pressure during rapid sequence induction for patients with acute cervical SCI; a gentle BURP maneuver could be utilized to improve glottic view.
Cervical Spine Distraction During Airway Instrumentation
Although neurologic impairment during airway manipulation is very uncommon, all airway maneuvers are associated with some degree of C-spine mobilization. However, this is based only on retrospective evidence and expert opinion. In patients with cervical cord injuries, it is generally better to use little or no anesthetic drugs, except as needed for safe conscious sedation and then to use fiberoptic bronchoscopy with good topicalization. This does require cooperation of the patient and if that is not possible or the intubation needs to be done in an emergency, direct laryngoscopy with in-line immobilization has not been associated with apparent neurologic complications, although its anatomic efficacy has been questioned. Notably, Lennarsen et al. reported in fresh cadavers that neither traction nor in-line immobilization prevented distraction nor angulation of the C4–5 injured cervical spine with laryngoscopy by Macintosh blade. Because concern persists about this approach producing or exacerbating cervical spine injury, it remains reserved for emergencies or for the uncooperative patient in whom general anesthesia with rapid intubation by laryngoscopy is the only viable option. In fact, direct laryngoscopy with manual in-line stabilization is the most commonly used technique for emergency endotracheal intubation in patients with acute cervical SCI, and it is recommended according to Advanced Trauma Life Support guidelines. In a survey of 122 physician anesthesiologists, indirect techniques were preferred for elective intubation of patients with cervical spine injury (CSI) and direct laryngoscopy was preferred for emergency intubation.
There have been many reports of various approaches to endotracheal intubation with various endpoints. Some are simply case reports indicating that a given technique worked in a given patient with cervical vertebral instability. However, a prospective randomized study that provides conclusive outcome data favoring a given approach remains unpublished. Nonetheless, there are some studies that used other reasonable surrogate endpoints that can provide an element of evidentiary support for some approaches.
Basic airway maneuvers such as mask ventilation, chin lift, and jaw thrust have the potential for displacement of the cervical spine even when a cervical collar is in place. The sniffing position entails flexion of the lower neck and extension of the head on the upper neck. In anesthetized normal patients, the majority of cervical movement during classic Macintosh laryngoscopy occurs at the occipitoatlantal and atlantoaxial articulations (C1–C2). Lower cervical segments are displaced only minimally with direct laryngoscopy. In cadaver studies with cervical instability, chin lift and jaw thrust were found to result in more C-spine displacement and increase in disc space than either oral or nasal intubation, but jaw thrust without chin lift or neck extension resulted in significantly reduced displacement.
Turkstra et al. in normal anesthetized patients carried out a fluoroscopic evaluation of neck extension produced by bag-valve-mask ventilation (BVM), Macintosh blade laryngoscopy, light wand, and GlideScope. They found minimal movement during BVM, with the most extension noted by Macintosh blade laryngoscopy. Both light wand and GlideScope produced much less neck extension than Macintosh blade laryngoscopy with the light wand taking an equivalent time and the GlideScope taking longer.
Rudolph et al. compared Macintosh laryngoscopy with a rigid Bonvil fiberoptic scope reporting less cervical spine motion with the rigid fiberoptic scope but it was not zero. Hastings et al. evaluating external extension without fluoroscopy reported very little neck extension with the Bullard laryngoscope compared with Miller and Macintosh blades.
Robitaille et al. reported that during intubation under general anesthesia with neuromuscular blockade and manual in-line stabilization, the use of GlideScope produced better glottic visualization but did not significantly decrease movement of the nonpathologic C-spine when compared with direct laryngoscopy.
Brimacombe et al. in cadavers with C3 instability compared the effects of facemask application, esophageal/tracheal Combitube, laryngoscopy guided intubation, flexible fiberoptic-guided intubation, intubating laryngeal mask airway (LMA), and maximal head/neck flexion and extension. Displacement was negligible with the fiberscope, moderate with LMA and with facemask/chin lift, worse with laryngoscopy, and worst with Combitube. Extension of the neck data as a percentage of the maximum seen with purposeful neck extension are presented in Fig. 24.14 .
In another fluoroscopic study of C-spine injured cadavers, Gerling et al. compared laryngoscope blades combined with in-line immobilization, reporting that the Miller blade produced less cervical vertebral displacement than the Macintosh blade and that manual immobilization is superior to cervical collar during intubation.
The use of video laryngoscopes such as GlideScope, C-MAC, McGrath, Airtraq, etc., has been associated with improvement in glottis visualization when manual in-line stabilization is used and probably less C-spine displacement. However, procedural time may be longer for inexperienced clinicians. Ruetzler et al. reported that in a simulated setting of difficult intubation, GlideScope and C-MAC were superior compared with conventional direct laryngoscopy. In a recent systematic review and meta-analysis of RCTs comparing any intubation device with the Macintosh laryngoscope in humans with C-spine immobilization, Suppan et al. reported that the Airtraq was associated with a significant reduction of intubation failure at the first attempt, a higher rate of Cormack-Lehane grade 1, reduction of operational time, and oropharyngeal complications. Brück et al. compared C-MAC and GlideScope in patients with C-spine disorders and reported that both devices provide an excellent glottic view, but tracheal intubation was more successful on the first attempt with the GlideScope. Yumul et al. compared the C-MAC video laryngoscope with the standard fiberoptic scope with an eye piece for intubation in patients with C-spine immobilization. The glottic view at the time of intubation did not differ significantly with the two devices; however, the C-MAC facilitated more rapid tracheal intubation compared with the fiberoptic scope. The peak heart rate response following insertion of the tracheal tube was also lesser in the C-MAC group. Holmes et al. reported that among patients with acute CSI at a high-volume academic trauma center, video laryngoscopy was the most commonly used initial intubation technique. Awake fiberoptic bronchoscopy (FOB) and direct laryngoscopy were performed infrequently and no cases of neurological deterioration secondary to airway management occurred with any method.
Fiberoptic intubation, either asleep or awake, causes little motion of the cervical spine. However, coughing and/or gagging can occur during awake intubation if not adequately topically anesthetized, resulting in motion of the injured spine. In fact, fiberoptic scope in an emergency situation with providers who are inexperienced with this device has a high failure rate and can be very challenging if the patient is uncooperative or with excessive blood or secretions in the airway.
In summary, airway management in the presence of acute CSI is probably the most challenging scenario for the anesthesiologist. Traditionally, awake fiberoptic has been recommended because it limits cervical spine motion during tracheal intubation and allows neurological examination after the procedure. However, in the real world the brain-injured patient typically is not cooperative such that the agitation that might ensue with fiberoptic attempts can produce unacceptable patient-induced neck movement. This then results in the need for increasing amounts of sedation, in the context of a full stomach. This may or may not produce sufficient sedation while putting the patient in an unacceptable in-between state of pharmacologically depressed sensorium with persisting agitation and an unprotected airway. The situation may be enough of an emergency that the use of a fiberoptic technique, in a setting of developing hypoxemia, full stomach, possible airway injury with associated secretions, means the planned nonexpeditious intubation has an unacceptable risk of hypoxemia and/or aspiration. For these reasons, in such situations many clinicians advocate direct laryngoscopy with appropriate doses of CNS depressants and neuromuscular blockade with immobilization or in-line traction. If a difficult airway problem arises in this situation, there should be no extreme laryngoscopy attempts that will clearly risk quadriplegia. Rather the algorithm should move quickly to LMA to at least establish oxygenation followed by securing the airway via LMA and then by intubation, or by tracheostomy.
Currently, with the widespread availability of video laryngoscopy, fiberoptic use is declining dramatically. Assuming care is taken to limit neck movement, providers should use the intubation technique with which they have the most experience and skill. In applying these data to an individual patient, considerable judgement may be needed. Often the best evidence-based method may not fit your individual patient situation. Nonetheless, video laryngoscopy of the airway with intubation of the trachea seems to be the new gold standard and certainly should be associated with near zero risk of exacerbation of spinal cord injury.
PaO 2 Physiology
The conceptual relationship of CBF to PaO 2 is depicted in Fig. 24.15 . Hypoxemia, usually below a PaO 2 of approximately 50–60 mmHg, is associated with vasodilation. At the opposite extreme, Floyd et al. in a group of healthy volunteers demonstrated the vasoconstrictive effect of hyperoxemia and its accompanying hypocapnea ( Fig. 24.16 ). Nakajima et al. evaluated this phenomenon in patients with cerebrovascular disease, finding that areas of the brain with impaired cerebrovascular reserve were not adversely affected by hyperoxia.
The optimal PaO 2 to seek in a brain injured patient is presently unclear. There are data that support hyperoxic therapy along with data that suggest such an approach is deleterious. In addition, the bedside decision about PaO 2 management is further coupled to cerebrovascular reserve issues previously discussed. Thus, a low PaO 2 that would normally be tolerated through vasodilation may not be so well tolerated if vasodilatory reserve is compromised with, for example, carotid occlusion, brain edema, or anemia.
Fiskum et al. among others have reported in laboratory studies that hyperoxic therapy promotes generation of free radicals and that such oxidative stress causes mitochondrial injury, which will act to impair neurologic recovery. This notion from in vitro considerations is supported by in vivo studies in rodents and dogs, which demonstrated worse neurologic outcomes when hyperoxia is used before or after an ischemic insult.
Conversely, with the advent of reports supporting the feasibility and reliability of brain tissue PO 2 monitoring, data regarding increased brain tissue oxygen monitoring show that normoxemic therapy, in the context of cerebral hypoxia, may promote ischemic injury. Indeed, many recent reports have described in nonrandomized retrospective and prospective series of both traumatic and SAH patients that brain tissue PO 2 less than 20–30 mmHg is associated with worsened neurologic outcome ( Fig. 24.17 ). Notably, however, these studies did not examine effects of hyperoxia, which is the negative situation identified by Fiskum et al. These studies point out the value of avoiding tissue hypoxia, perhaps at the cost of systemic hyperoxia but not intracranial hyperoxia. However, one side observation made by these studies is the impact of avoiding hyperoxia because brain tissue oxygen monitoring allows the provision of minimal FiO 2 that permits the optimal (not too high not too low) brain tissue oxygen level.
Given these potentially conflicting therapeutic priorities, it seems that the most sensible approach at this time is as follows: In the presence of a brain tissue PO 2 monitor, adjust physiologic parameters to keep PbrO 2 > 20 mmHg. This may entail use of FiO 2 > 60% with concomitant risk of pulmonary oxygen toxicity. It seems that this risk can be incurred for 1 to 2 days but that continued dependence on pulmonary toxic oxygen concentrations should produce a time-dependent increase in pressure on the caregivers to decrease FiO2, even if this means use of higher airway pressure or allowing PbrO 2 to decrease after a few days.
In the absence of a PbrO 2 monitor, the clinician is left basing therapy on assumptions about brain oxygenation. If it is felt that many brain areas are well perfused and at risk for hyperoxia, pulmonary management should aim for PaO 2 just sufficient to produce SaO 2 > 95%. Conversely, if there is elevated ICP and/or areas of brain hypoperfusion then a reasonable empiric approach would be to utilize an FiO 2 of 0.60. This will maximize PaO 2 and PbrO 2 but without significant risk of acute pulmonary injury.
PEEP and Intracranial Hypertension
Positive end expiratory pressure (PEEP) can increase ICP ( Fig. 24.18 ), and this relationship has been the subject of clinical research for decades. Two mechanisms can be posited. The first is through impedance of cerebral venous return to increase cerebral venous pressure and ICP. The second is through decreased blood pressure and reflex increase in cerebral blood volume to increase ICP. Huseby’s data suggest that cerebral venous effects only occur with very high PEEP.
Frost demonstrated that applying PEEP between 5 and 12 cm H 2 O (and even transiently up to 40 cm H 2 O improved arterial oxygenation without increasing ICP. Shapiro et al. demonstrated increases in ICP in head-injured humans with intracranial hypertension with application of PEEP (see Fig. 24.18 ). Examination of their data indicates that the most profound decreases in CPP occurred in patients with PEEP-induced decrements in mean arterial pressure consistent with the notion put forth by Rosner that decreases in blood pressure increase CBV to increase ICP Aidinis et al. Studies in cats confirmed these observations in a more controlled setting. In addition, they assessed the role of pulmonary compliance, finding that decreased pulmonary compliance with oleic acid injections results in less of an effect of PEEP to increase ICP. Such observations indicate in situations where PEEP is likely to be needed, often accompanied by decrements in pulmonary compliance, that any adverse effects on ICP are less likely to manifest. This may be related to observations that hemodynamic effects of PEEP are less apparent with noncompliant lungs such that hypotensive-mediated increases in CBV do not occur.
Historically, use of high PEEP has been associated with potential for ICP elevation; however, the intuitive notion that PEEP increases cerebral venous pressure to increase ICP is not as straightforward as some may indicate. For PEEP to increase cerebral venous pressure to levels that will increase ICP, the cerebral venous pressure must at least equal the ICP. Thus, the higher the ICP, the higher PEEP must be in order to have such a direct hydraulic effect on ICP. This concept was proven by Huseby et al. in dog studies in which PEEP was increased progressively with different starting levels of ICP ( Fig. 24.19 ). It is important to note that they prevented PEEP-induced decrements in blood pressure, thus avoiding any reflex increases in cerebral blood volume. They suggested a hydraulic model to conceptualize this better ( Fig. 24.20 ). Thus, for example, if all of a 10 cm H 2 O PEEP application was transmitted to the cerebral vasculature, which is unlikely given the decreased pulmonary compliance associated with the need for such PEEP, then ICP will only be affected if it is < 10 cm H 2 O (7.7 mmHg) with it increasing to a level no higher than the applied PEEP. Such observations are consistent with the notion that there is a Starling resister regulating cerebral venous outflow.