Head Injury

Published on 07/03/2015 by admin

Filed under Critical Care Medicine

Last modified 07/03/2015

Print this page

rate 1 star rate 2 star rate 3 star rate 4 star rate 5 star
Your rating: none, Average: 5 (1 votes)

This article have been viewed 6280 times


Head Injury


In the context of critical care medicine, the management of severe head injuries remains the Achilles’ heel of neurosurgery, as borne out by our own observations over three decades of practice at academic centers. Although revolutionary advances now allow imaging of the brain and spinal cord with greatly improved speed and anatomic detail, translation of this resource into better outcomes has been disappointing. The search for a neuroprotective agent that can consistently prevent the deadly cascade of events that leads to irreparable brain damage remains for future investigators as of this writing.

Key to optimizing clinical outcomes has been the recognition that the time from injury to surgery must be as short as possible to minimize the secondary manifestations of serious brain trauma. Other valuable lessons from clinical experience reported in the literature are that glucocorticoids have no significant therapeutic benefit in managing severe head injuries and that hyperventilation, very effective in reducing high intracranial pressure (ICP), must be used judiciously; otherwise, it may become more harmful than helpful.

It also has been demonstrated that the management of serious brain trauma requires specialty teams consisting of experienced neurosurgeons, traumatologists, and intensivists (specifically, neurointensivists) working together in a dedicated unit equipped and staffed to optimally care for these patients. In the absence of such facilities, outcomes suffer.1-3

Finally, the introduction of a simple-to-use and highly predictable neurologic status scoring system, the Glasgow Coma Scale (GCS), has enabled meaningful multidisciplinary communication from the accident scene through intensive care and beyond, resulting in efficient and timely triage and treatment of these patients in tenuous conditions.


The Centers for Disease Control and Prevention report the number of patients sustaining a traumatic brain injury (TBI) to be at least 1.7 million per year in the United States. Approximately 80% of patients are released from the emergency department, 275,000 are admitted, and 52,000 die. Nearly a third of all traumatic deaths in the U.S. involve a head injury. The ages most likely to sustain a TBI are children up to 4 years old and adults over 64, with adults greater than 74 years old having the highest rates of hospitalization and death associated with head injury.4 Teenagers and young adults continue to be at high risk for sustaining TBI, but the range has decreased over the past 20 years from between 15 and 24 to between 15 and 194,5. The incidence of head injury has been shown to be inversely proportional to economic status,6 and has remained up to three times more common in males in every age group over the past several decades.4,7,8 In the United States, falls are the most common cause of TBI and are responsible for half of the head-injured children up to 14 years of age and 60% of the adults over the age of 64.4 Although the frequency of head injury related to motor vehicle crashes has decreased over the years to a little less than 20%, these crashes are the second leading cause of head injury in all age groups and result in the largest percentage of TBI-related deaths.4,9 The incidence of TBI worldwide is difficult to quantify but is estimated to be at least 10 million people a year, with traffic accidents accounting for more than half, and falls being the second leading cause;10 this is the opposite distribution seen in the United States4. Violence is estimated to cause about 10% of head injuries throughout the world,10 as well as in the U.S.4, but there is significant variability in the specific mechanisms. For example, between 1997 and 2007 in the U.S., firearms were related to just over a third of all lethal TBIs, compared to a single hospital in South Africa that reported its experience with more than 300 stab wounds to the brain.11

Diagnostic Approach

The presenting neurologic condition of patients with head injuries is the primary factor in determining the initial management and prognosis. When a detailed history is unavailable, it is important to keep in mind that loss of consciousness may have preceded and in fact caused the traumatic event, such as aneurysmal subarachnoid hemorrhage (SAH), hypoglycemia, intoxication, or syncope. Level of consciousness is one of the most important neurologic considerations in managing a head-injured patient. Neurosurgical patients generally have an alteration in level of consciousness from either brainstem or bilateral cerebral hemispheric involvement. This may be secondary to poor perfusion due to high ICP or brainstem compression, both of which require neurosurgical intervention. Therefore, an accurate tool to measure level of consciousness is essential. Many clinical assessment tools are available for use in the critical care setting.12-24

Glasgow Coma Scale

Ideally, using the GCS, a nurse, medical student, paramedic, physician assistant, intensivist, trauma surgeon, emergency medicine physician, neurologist, and neurosurgeon all should obtain the same score when assessing a patient. The GCS system is not perfect, but it has proved itself to be, overall, a practical, straightforward grading system that can be used by health professionals in various fields and at different levels to produce reliable results.25-30 The GCS has become the most widely used assessment tool and is considered the gold standard for the evaluation of patients with head injuries.12

The GCS is a measure of level of consciousness and does not take into account focal deficits. It is based on eye opening (1-4), verbal response (1-5), and motor response (1-6) (Table 66.1). A patient with a normal level of consciousness should have the highest possible score of 15. The lowest score possible is 3, not 0 as might be expected. An intubated patient technically gets a 1 for verbal response and is assigned 1T (for Tube) for the verbal score. It is important to identify the score for each observed variable tested. For example, a typical score for an intubated patient with decorticate posturing probably would be E1 + V1T + M3 = 5T (where E denotes eye opening, V is verbal response, and M is motor response). It also is important to realize the shortcoming of this system in evaluating patients with dementia or aphasia. A point to keep in mind is that the GCS attempts to put a numeric value on level of consciousness, with 15 being representative of normal. If a demented patient from a nursing home presents with a normal level of consciousness after a fall but doesn’t know what day it is, the GCS would be calculated as E4 + V4 + M6 = 14. A GCS score of 14 describes a patient with an abnormal level of consciousness but fails to accurately communicate this particular patient’s condition. Likewise, clinicians must use caution when applying the GCS to patients with aphasia. Conversely, patients with profound focal deficits and a normal level of consciousness should have a GCS score of 15. Another problematic case is that of a quadriplegic patient with a normal level of consciousness who is able to blink the eyes or stick out the tongue upon command, thereby giving a top score of 6 on the motor assessment.

Table 66.1

Glasgow Coma Scale

Component Score
Eye Opening  
Spontaneously 4
To voice 3
To pain 2
No response 1
Oriented 5
Disoriented 4
Inappropriate 3
Incomprehensible sounds 2
No response 1
Following commands 6
Localizing to pain 5
Withdrawing to pain 4
Abnormal flexion 3
Abnormal extension 2
No response 1

The GCS has been used to predict outcome.31 The motor score itself has predictive value as well.32 The GCS score should be calculated after hemodynamic and pulmonary resuscitation and without sedatives or muscle relaxants.33

Aggressive implementation of early sedation and intubation in severely head-injured patients compromises the ability to determine an accurate GCS score.34

Computed Tomography Classification: Marshall Classification

A useful classification of brain injuries by findings on computed tomography (CT) has been devised by Marshall and colleagues. This CT classification, presented in Table 66.2, can serve as a guide in describing scans and has strong predictive power.34,35

Table 66.2

Marshall Computed Tomography Classification

Category Definition
Diffuse injury I (no visible pathology) No visible intracranial pathology seen on computed tomography scan
Diffuse injury II Cisterns are present with midline shift of 0-5 mm or lesion densities present no high- or mixed-density lesion > 25 cm3 may include bone fragments and foreign bodies
Diffuse injury III (swelling) Cisterns compressed or absent with midline shift of 0-5 mm; no high- or mixed-density lesion > 25 cm3
Diffuse injury IV (shift) Midline shift > 5 mm; no high or mixed-density lesion > 25 cm3
Evacuated mass lesion Any lesion surgically evacuated
Nonevacuated mass lesion High- or mixed-density lesion > 25 cm3; not surgically evacuated

From Marshall LF, Marshall SB, Klauber MR, et al: A new classification of head injury based on computerized tomography. J Neurosurg 1991;75:S14-S20.

Maas and associates have proposed a scoring system with better predictive power (Table 66.3), with a sum total adjusted to be consistent with the GCS.34 Table 66.4 presents a mortality prediction chart based on this classification.

Table 66.3

Prognostic Score Chart for Probability of Mortality in Patients with Severe or Moderate Traumatic Brain Injury by Computed Tomography Characteristics

Predictor Score
Basal Cisterns  
Normal 0
Compressed 1
Absent 2
Midline Shift  
No shift or shift < 5 mm 0
Shift > 5 mm 1
Epidural Mass Lesion  
Present 0
Absent 1
Intraventricular Blood or tSAH  
Absent 0
Present 1
Sum Score* +1

*The sum score can be used to obtain the predicted probability of mortality (see Table 66.4). The authors chose to add plus 1 to make the grading numerically consistent with the grading of the motor score of the GCS and with the Marshall CT classification. tSAH, traumatic subarachnoid hemorrhage.

This scoring system is best used as one piece of data in the overall clinical assessment in considering a patient’s prognosis.36

Prediction of Outcomes

Severe brain injury outcomes can be accurately predicted using four variables: age, GCS score on admission (especially motor score), CT characteristics, and presence of ischemic and hemodynamic secondary insults. Using these four predictors, numerous investigators have used statistical modeling techniques to predict up to 80% or more of outcomes.37

Head injuries generally are classified as mild (GCS score of 13 to 15), moderate (GCS score of 9 to 12), or severe,3-8 although the evidence may be sufficient to include a GCS score of 13 in the moderate category.38

CT scanning is the diagnostic imaging study of choice in trauma. Plain x-ray films of the skull are rarely indicated as a screening study for head-injured patients. Outcomes with the strategy of obtaining CT scans in all head-injured patients are superior to those with other management strategies. The incidence of missed surgical lesions in mildly head-injured patients by CT scanning is 0.028%.39-42

Primary Head Injury

Primary head injury can be defined as the damage that occurs at the moment of impact and can take the form of skull fractures, surface contusions and lacerations, diffuse axonal injury, or diffuse vascular injury.42 Some authors include the contusions and hematomas that form as a direct and immediate effect of the impact as part of the primary injury as well.43 These hemorrhages may manifest with significant mass effect requiring surgical intervention.

Emergent neurosurgical intervention is designed to prevent permanent damage to the central nervous system (CNS) in general and to the reticular activating system in particular. Located in the brainstem, the reticular activating system is what allows a meaningful, awake condition; if it is destroyed, the patient will be vegetative. The surgical management of head-injured patients typically is driven by the presence of pathologic masses (hematomas) in anatomic spaces. These spaces are best described by reviewing the anatomic layers.

In virtually all cases of head injury seen in clinical practice, the pathogenic mechanism is an impact. The first layer involved is the scalp, the layers of which are best remembered by the pneumonic SCALP: S for skin, C for connective tissue, A for galea aponeurosis, L for the loose connective tissue layer, and P for periosteum. The scalp is the thickest skin in the body and absorbs some of the energy delivered to the head during impact. The scalp has a rich blood supply with relatively large blood vessels situated in the space where the connective tissue meets the galea. A large laceration has the potential for heavy blood loss and needs to be managed with local, direct pressure until surgical repair can be accomplished. Care must be taken in applying pressure if there is an underlying depressed skull fracture or, especially, a skull defect as is seen in missile injuries. It should be noted that an unattended, serious scalp laceration does have the potential for blood loss to the point of hemodynamic instability, but this is the only exception to the rule that an adult cannot lose enough blood from any intracranial hemorrhage to cause hypovolemic shock and hemodynamic instability. Blood clots within the scalp usually are limited to the loose connective tissue layer and are referred to as subgaleal hematomas, which may be a feature of massive injuries or a sign of coagulopathy but rarely require treatment other than direct pressure if active bleeding is suspected. A cephalohematoma is a blood clot that expands in the potential space between the periosteum and the skull; this lesion is limited to neonates.

Skull Fractures

The next layer involved in trauma is the skull, composed of outer and inner tables with the intervening vascular diploic space. Skull fractures are best considered as involving the cranial vault or the skull base. Cranial vault fractures are further divided into linear or depressed and open or closed. In general, closed, linear fractures of the cranial vault do not require any specific treatment.

Basilar skull fractures typically are linear and involve the anterior cranial base and the petrous part of the temporal bone. These fractures have the potential for an associated dural laceration adjacent to potentially contaminated paranasal sinuses, or the external ear canal if the tympanic membrane is disrupted. This allows the potential for a cerebrospinal fluid (CSF) fistula to develop, as well as meningitis. Clinical signs of a fracture of the petrous portion of the temporal bone include hemotympanum with or without tympanic membrane disruption, hearing loss, CSF otorrhea, and Battle sign. The cranial nerves that course through the temporal bone include the facial, acoustic, and vestibular nerves; therefore, associated vestibular dysfunction or facial weakness may be noted (Fig. 66.1). Anterior cranial base fractures may be associated with “raccoon eyes,” anosmia, and CSF rhinorrhea. It is important to remember that when a fracture of the floor of the anterior cranial base is suspected, a nasogastric tube should not be inserted, because of the risk of intracranial penetration (Fig. 66.2).

The use of prophylactic antibiotics in basilar skull fractures has been debated but generally is not recommended. When CSF leak does occur, the break in the tissue usually will heal with conservative treatment over the course of a week, unless the defect is very large or a bony spicule is identified. Keeping the patient’s head elevated and possibly using external CSF diversion after a few days are reasonable nonoperative methods; however, if the leak persists, surgical repair is indicated.44

Most depressed skull fractures occur in men younger than 30 years of age.45 Despite very little literature to support any particular management strategy, closed depressed skull fractures generally are operated on if the extent of depression is greater than the full thickness of the adjacent skull, with the theoretical benefits of improved cosmetic result, a decrease in late-onset posttraumatic epilepsy, and a reduction in the incidence of persistent neurologic deficit.46 Most depressed skull fractures, however, are open45 (Fig. 66.3), meaning that a scalp laceration with galeal disruption overlying the fracture is present.

Open depressed skull fractures may be associated with significant morbidity and mortality.45,47 Infection rates are reported to be between 1.9% and 10.6%,4650 with neurologic morbidity and mortality rates of approximately 11% each46 and an incidence of late epilepsy up to 15%.51

By convention, open depressed cranial vault fractures are treated surgically, with debridement and elevation, primarily to attempt to decrease the incidence of infection. Open depressed cranial fractures may be treated nonoperatively if clinical and radiographic examination reveals no evidence of dural penetration, significant intracranial hematoma, depression greater than 1 cm, frontal sinus involvement, gross cosmetic deformity, wound infection, pneumocephalus, or gross wound contamination.46

A special type of depressed fracture is fracture of the frontal sinus. Depression of the fracture may require surgery to prevent CNS infection and CSF leak (Fig. 66.4). Another special type of depressed fracture involves a major dural venous sinus (see Fig. 66.4). Under these circumstances, the risks associated with surgery are increased, and elevation of the fracture fragment generally is reserved for significant compromise of venous drainage.44

Epidural Hematoma

The next anatomic layer after the skull is the dura, which is the periosteum of the inner table of the skull and, as such, is tightly adherent to the bone. The potential space between the inner table of the skull and the dura is the epidural space. Epidural hematomas (EDHs) form in this potential space between the skull and the dura. The most common mechanism for the development of an EDH is a motor vehicle crash (in 53% of the cases), followed by falls (in 30%) and assault (in 8%).52-60 Typically, bleeding results from damage to the middle meningeal artery, but an EDH also can occur from injury to the middle meningeal vein, diploic veins, or the venous sinuses.60

EDHs generally occur in the temporal and temporoparietal regions (Fig. 66.5).52,57,6164 A useful way to describe the development of an EDH is to follow the events that take place when a patient presents with an initial lucid interval following brief loss of consciousness after head trauma. A lucid interval occurs when a patient initially is rendered unconscious from a concussive head injury that causes a linear skull fracture involving the middle meningeal artery or one of its branches. The middle meningeal artery is a dural vessel that runs half in the dura and half through the groove in the inner table of the skull. When a fracture extends across the bony groove of the artery, it will tear and bleed. Because the dura is tightly adherent to the inner table of the skull, significant force is needed to push the dura off the inner table. An arterial hemorrhage generally has enough pressure to strip the dura off the bone, converting the potential epidural space into a mass. The bony attachment of the dura becomes progressively stronger with age; therefore, the dura of a younger patient requires less force to push it off the bone than would be required in an older patient. Not surprisingly, the mean age of patients with EDH is between 20 and 30 years of age,5355,6370 and EDHs are unusual in patients older than 50.60

The growing epidural mass commonly compresses the anterior temporal lobe, which usually will not cause a major detectable neurologic deficit early in the course in the emergency department. During this so-called lucid component of the lucid interval, the patient regains consciousness from the concussive injury while the EDH is expanding and compressing the relatively silent anterior temporal lobe. The medial part of the temporal lobe, the uncas, lies just lateral to the brainstem at the level of the third cranial nerve (oculomotor nerve), which runs alongside the tentorial edge. When the EDH enlarges to the point at which it pushes the uncas of the temporal lobe over the tentorial edge, rapid development of a third nerve palsy, along with brainstem compression, is possible—which, at this level, will cause a contralateral hemiparesis from direct compression of the cerebral peduncle and a decrease in level of consciousness from the effect on the reticular activating system. The classic signs at this point are coma with a dilated pupil ipsilateral to the EDH and a contralateral hemiparesis. Occasionally, instead of directly compressing the ipsilateral cerebral peduncle, the herniating uncas can shift the brainstem into the contralateral tentorial edge, a relatively sharp rigid structure. This can damage the brainstem on the side opposite the EDH and cause a hemiparesis ipsilateral to the side of the hematoma. This is known as Kernohan’s notch phenomenon.

Taking a general view, the patient goes from unconscious at the time of impact from a concussive injury, to awake, to unconscious again secondary to brainstem compression—hence the term lucid interval. The lucid interval is observed in close to half of the patients undergoing surgery for EDH.*

Pupillary abnormalities occur in approximately 20% to 30% of patients with surgical EDH.52,54,59,60,67 Cranial fractures are present in between 70% and 95% of the cases.54,60,68,70,74,75 Associated intracranial lesions are found in 30% to 50% of adults with surgical EDHs, and subdural or parenchymal lesions in association with EDH lower the chance of a good outcome.60

The overall mortality rate (for all ages and GCS scores) is approximately 10%. The time lapse between the onset of pupillary abnormalities and surgery determines outcome,61,66,80 but the single most important predictor of outcome in patients operated on for EDH is the GCS score on admission and before surgery.52,53,55,61,8183

Not all EDHs, however, require surgery. No prospective randomized trials have been conducted to compare surgical treatment with nonoperative management, nor should there be. Available data describe nonsurgical management in selected cases. In one study, approximately 10% of the total number of EDHs were treated nonoperatively. All were conscious with a GCS score greater than 11 and a midline shift on CT scan of less than 10 mm. Not one of these nonoperative hematomas was in the temporal region.84

Another study reported findings in a group of 57 selected patients treated nonoperatively with an initial GCS score of 10 or higher, with maximum hematoma thickness less than 13 mm, with five clots located in the temporal region, but only one patient had a midline shift on CT.53

An interesting approach for dealing with thin, acute EDHs in an early stage has been reported in a small, isolated series of patients using endovascular techniques to occlude the middle meningeal artery.85 Surgical decision making for an acute EDH is based on GCS score, pupillary findings, comorbid conditions, CT findings, age, and, in delayed decisions, ICP. Guidelines for the surgical management of acute EDH published in 2006 recommend surgical evacuation of an EDH less than 30 mL in volume. Smaller hematomas in patients with a GCS score greater than 8 and without focal deficits, clot thickness less than 15 mm, and less than 5 mm of midline shift may be considered for nonoperative management. Close neurologic monitoring and serial CT scanning are essential. Of importance, a temporal location for an EDH is associated with failure of nonoperative management and should lower the threshold for surgery.60

Because time between neurologic deterioration and surgery is critical, the question of whether a patient with an acute EDH should receive treatment at the nearest hospital or should be transferred to a trauma center is important. The issue is whether or not a non-neurosurgeon should operate on a patient deteriorating from an acute EDH as a true emergency. One suboptimally controlled study reported worse outcomes in a small group of patients who underwent emergency operations by non-neurosurgeons and attributed this mainly to the technical inadequacy of the operation.59,60 Other studies have documented worse outcomes in patients transferred from outlying hospitals for surgery.52,73 The take-home message is that EDHs are extra-axial hemorrhages located adjacent to the brainstem that can represent a true neurosurgical emergency, and that rapid, competent decompression makes a difference.

Subdural Hematoma

Proceeding deep to the skull, the next space is the subdural space. Unlike the potential epidural space, the subdural space is a real space. It is a compartment that follows the contour of the brain, which is how it appears on a CT scan (Fig. 66.6). Anatomically, bridging veins course through the subdural space from the cerebral hemispheres to the superior sagittal sinus. As the brain accelerates within the skull after impact, these veins can stretch and tear.86

The source of bleeding also can be from a cortical artery87,88 or vein. These hematomas typically form over the surface or between the cerebral hemispheres, along the tentorium, or between the temporal lobe and base of the skull.

An acute subdural hematoma (SDH), as defined by Bullock and colleagues, appears within 14 days of injury,60 although some authors consider an SDH to be subacute when signs and symptoms develop between 3 and 20 days after trauma.89

Acute SDHs consist of clotted blood, which is fibrous and often adherent to adjacent tissue. The blood remains clotted for several days. After this time, the clot gradually and progressively lyses, resulting in a mixture of clot and fluid. After several weeks, the clot is liquefied and becomes a chronic SDH. Chronic SDHs may manifest weeks or months after what may have been mild or insignificant trauma. Chronic SDHs occur more often in elderly patients and individuals who have more intracranial space because of cortical atrophy. A membrane often surrounds these hematomas, and the collection of fluid may slowly grow in size because of repeated small bleeds or accumulation of fluid transudate from the membrane.86

The incidence of acute SDHs is approximately 20% in severe traumatic brain injury,60,9093 and the mean age is between 31 and 47 years, with most of the patients being men.60,9497 The mechanism of injury differs between age groups, with patients older than 65 more commonly presenting after a fall, and younger patients being involved more often in a motor vehicle crash,60,98100 which, among comatose patients with acute SDH, is the most common mechanism of injury.60,91,93,101

Between 37% and 80% of patients with acute SDH present with a GCS score of 8 or lower,60,65,71,94,97,102 and the overall mortality rate is between 40% and 60%.* Mortality rates among comatose patients requiring surgery are somewhat higher, with reported rates between 57% and 68%. Acute subdural hematomas frequently are associated with other brain injuries,97,102 which is one of the reasons why the mortality rate is relatively high.

A simple acute SDH is a subdural, extra-axial hematoma without any other associated brain injury (Fig. 66.7). A complex SDH is associated with parenchymal injury (Fig. 66.8).107,108 Fewer than half of acute SDHs requiring surgery are isolated, simple lesions.60,97,102

One useful method to identify the presence of associated injury is to measure the thickness of the subdural clot relative to the amount of brain shift on the CT scan. If the amount of shift is directly proportional to the thickness of the extra-axial clot, the injury is likely to be simple and only the brain compressed. However, if the amount of shift is more than expected as indicated by the size of the hematoma, then additional parenchymal brain injury probably is present under the clot, causing additive mass effect. Not surprisingly, the mortality for complex injuries is higher than that for simple SDHs.107,108

In addition to the obvious compressive effects that an SDH generates, the available evidence points to a direct toxic effect of the blood itself on the underlying cortex, thereby compounding the problem.109,110

Age is an important factor in acute SDHs. There is a significant increase in poor outcome among patients older than 60 years of age with severe head injury in general.93,98101,111,112 Older patients with an acute SDH and a low GCS score do especially poorly.98101,111,112

The decision for immediate surgery is dependent on GCS score, age, pupillary examination, comorbidities, CT findings, and salvageability with respect to the patient’s level of injury. In addition to all of these factors, decision making for delayed surgery is dependent on clinical course and ICP.60

On the basis of a contemporary search of the literature, Bullock and colleagues found that CT parameters of a midline shift greater than 5 mm and a clot thickness greater than 10 mm were independent factors requiring surgery in salvageable patients. They also identified a select group of comatose patients with smaller SDHs who could be managed nonoperatively if the patients remained neurologically stable with normal pupils and ICP of 20 mm or less.60

Traumatic Subarachnoid Hemorrhage

The subarachnoid space is a CSF-filled compartment within which are the major cerebral blood vessels. The CSF within the subarachnoid space fills the basal cisterns and interdigitates into the cortical sulci. Traumatic SAH can be caused by bleeding of cortical arteries, veins, or brain surface cerebral contusions.113

In contrast with hemorrhages in other locations, SAH is not a discrete surgical clot that requires evacuation. Trauma is the most common cause of SAH; a ruptured aneurysm is the most common cause of a spontaneous SAH. Aneurysmal SAH generally involves the suprasellar cistern, where the circle of Willis lies. Traumatic SAH can be found as small-volume hemorrhages in the sylvian fissures and especially in the interpeduncular cistern. Traumatic SAH also commonly involves the cerebral convexity and can fill cortical sulci, which can sometimes mimic sulcal effacement114 (Fig. 66.9A).

Occasionally, the distribution of the hemorrhage may be hard to differentiate from an aneurysmal hemorrhage. Under this circumstance, a vascular study should be obtained (Fig. 66.9B). Cerebral vasospasm following aneurysmal SAH is a common, well-described problem with an unclear pathogenesis. Increasing evidence suggests that SAH from trauma also may cause clinically significant cerebral vasospasm, which may be responsible for ischemia and infarction.115,116

Clearly the incidence of clinically relevant cerebral vasospasm is less than what is seen in aneurysmal SAH; however, posttraumatic SAH may cause ischemia during the acute as well as the delayed phase.115,117

In severe nonpenetrating head injury, the degree of SAH can be predictive of outcome,118,119 and although traumatic SAH can be an isolated finding, it is commonly associated with other intracranial injuries.120

Traumatic SAH can be a marker of severe primary injury and the amount of blood can also be an independent predictor of the development and progression of intraparenchymal contusions113 (Fig. 66.9C).

Intraparenchymal Contusions and Hematomas

Traumatic parenchymal mass lesions are common and are reported in 13% to 35% of severe traumatic brain injury.121-127 Contusions consist of heterogeneous areas of necrosis, pulping, infarction, hemorrhage, and edema.89,128,129 Hemorrhagic contusions are mixtures of blood and edematous cerebral parenchyma that also have a heterogeneous appearance on CT.89 Contusions commonly occur in the frontal and temporal lobes both at the poles and on the inferior surfaces as a result of contact with the rough, bony skull base130 (Fig. 66.10A).

Contusions typically involve the crests of the gyri but in more severe injury may extend into the substance of the white matter. If the pia-arachnoid is torn, contusions are classified as cortical lacerations.130-132

Contusions represent one end of a spectrum of injury on which hematomas, which are well-defined, homogeneous collections of blood, are the other end (Fig. 66.10B). The amount of energy delivered may have only been enough to cause failure of small vessels, resulting in contusions. If more energy is delivered, failure of larger vessels may occur, resulting in hematomas. Alternatively, the hemorrhagic component of a contusion may continue to bleed and coalesce into a more discrete hematoma. In general, the most common traumatic parenchymal lesions are contusions,89 and they tend to evolve.122,126,127,133,134

Risk factors for the progression of intraparenchymal hemorrhage include the presence of SAH (see Fig. 66.9C), and subdural hematoma, as well as the size of the clot on the initial CT scan. Enlargement of contusions occurs approximately 40% of the time, which justifies early follow-up CT scanning.135 In addition to the problem of progressive enlargement of existing contusions is a phenomenon of delayed appearance of traumatic intracerebral contusions (DITCH). DITCH is defined as a new contusion identified on a CT scan in an area of brain that was normal on the admission CT scan. These delayed hemorrhages are reported to occur in up to 7% of patients with severe head injuries.90

Contusions can be subdivided into two groups. Coup contusions occur in the brain tissue under the impact site and usually are associated with an acceleration injury. Contrecoup contusions are located away from the point of impact and usually are associated with a deceleration injury.130,131,136138

Understanding the biomechanics of how a contrecoup contusion develops is useful. In general, biologic tissues tolerate strain better if they are deformed slowly rather than quickly. For example, a 150-mL hematoma with accumulation of blood over minutes is likely to be lethal, whereas a 150-mL slow-growing meningioma may have no appreciable effect on the patient’s level of consciousness. In trauma, the cause of tissue damage may be any of three types of induced strains: compression, tension, and shear. The skull, brain, and blood vessels will tolerate compression better than tension, and tension better than shear. These strains are induced by contact or inertia (relative to acceleration-deceleration), or both. Contact injuries are a result of impact, which may cause inward deformation of the skull with local effects, and of shock waves, which can produce remote effects. As a consequence of contact the head is set in motion, which leads to inertial injury. Inertial injuries may cause damage by differential acceleration of the skull and brain. In addition, acceleration-deceleration can independently produce strains in the brain itself. The two clinically relevant types of acceleration are translational (when the brain moves in a straight line) and angular. Most injuries are a combination of both.139

These effects are best illustrated by a case example of a 37-year-old man who falls from a height onto the back of his head and presents with a deteriorating level of consciousness. His CT scan (Fig. 66.11) reveals soft tissue swelling at the point of impact in the right occipital region with a small underlying hemorrhage, as well as a large, hemorrhagic contusion in the left frontal lobe. The mechanism involved is primarily translational (linear) deceleration. The skull stops suddenly as it hits the ground, but the brain continues to move toward the impact site, where compression is induced (Fig. 66.12).

As the brain is moving toward the inner table of the skull at the impact site, it also is moving away from the skull on the opposite side, creating regions of low pressure and tensile strains. Contributing to the extensive tissue damage in the left frontal lobe also may be movement of the brain across the rough surface of the anterior cranial base.89,139

The same amount of energy is delivered to the brain in nearly a straight line, with the compressive injury at the point of impact resulting in much less damage than the contrecoup injury caused by tension. In view of the dramatic differential susceptibilities of the brain and blood vessels to compressive and tensile strains induced in trauma, it is understandable that diffuse injuries can be caused by seemingly less violent trauma when the head undergoes an injury with a large component of angular acceleration-deceleration. This is the motion that can induce shearing strains, which the tissues tolerate poorly (see “Diffuse Axonal Injury” later on).

Surgical decision making is more straightforward for epidural and subdural types of hematomas, in which the extra-axial mass causing compression is simply on the surface of the brain. Contusions and hematomas, however, are intra-axial masses intimately associated with surrounding regions of brain that may be salvageable. It is one thing to surgically evacuate an extra-axial hematoma compressing the dominant frontal lobe, and quite another to operate on a large contusion within the dominant frontal lobe.

In 2006, guidelines were published after a thorough review of the relevant but scientifically weak literature. The reviewers reported that patients with parenchymal mass lesions and signs of progressive neurologic deterioration referable to the lesion, medically refractory intracranial hypertension, or signs of mass effect on CT scan should be treated operatively. Patients with GCS scores of 6 to 8 with frontal or temporal contusions greater than 20 cc in volume with midline shift of at least 5 mm or cisternal compression on CT scan and patients with any lesion greater than 50 cc in volume should receive operative treatment. Patients with parenchymal mass lesions who do not show evidence of neurologic compromise, whose ICP is controlled, and who demonstrate no significant signs of mass effect on CT scan may be managed nonoperatively with intensive monitoring and serial imaging. For patients with refractory intracranial hypertension and diffuse parenchymal injury with clinical and radiographic evidence for impending transtentorial herniation, the guidelines also recommended, as treatment options, subtemporal decompression, temporal lobectomy, or hemispheric decompressive craniectomy (Fig. 66.13). With regard to timing of surgery, the literature supports a bifrontal decompressive craniectomy within 48 hours of injury as a treatment option for patients with diffuse, medically refractory posttraumatic cerebral edema and resultant intracranial hypertension.127

Hypothalamic-Pituitary Injury

Injury to the pituitary and hypothalamus can complicate head injury. A prolonged loss of consciousness often is reported in patients with such injuries. Up to 80% of cases are associated with a fracture through the skull base. Injuries include (in decreasing order of frequency) pericapsular hemorrhage, hypothalamic infarction, posterior pituitary hemorrhage, anterior pituitary infarction, and rupture of the infundibulum.140-142 Clinically measurable decreases in pituitary hormone production are not seen until at least 75% of the gland is destroyed. Complete loss of production requires destruction of at least 90% of the gland.

Derangements in pituitary function can result in decreased production of any of the pituitary hormones. Of particular clinical significance is impairment of adrenocorticotropic hormone release that can result in secondary adrenal insufficiency (Addison’s disease). Physiologic stressors such as trauma, surgery, or infection can result in addisonian crisis, with potentially life-threatening results. Diabetes insipidus has been reported in a significant number of severe head injuries, with a high percentage of cases remaining permanent.143 Injuries to the pituitary often are not suspected in the acute period, and the diagnosis of hypopituitarism is therefore often delayed (Fig. 66.14).

Diffuse Axonal Injury

The term diffuse axonal injury (DAI) describes brain damage in a group of patients who become immediately unconscious or go into coma at the time of the head trauma.81 Depending on the severity of the injury, patients may have mild, moderate, or severe DAI. A large number of patients with severe head injury will have DAI. Because of the gradient acceleration difference of certain brain areas during primary impact, shearing forces at the gray-white junction, corpus callosum, or brainstem may occur.131,144 The result of these forces will be diffuse tearing of axons and small blood vessels. These lesions initially may be hemorrhagic, but as they become chronic, shrinking, softening, and scarring ensue, often with cyst formation.42,145 The lesions usually are microscopic but can be large and macroscopic. Although small, focal, petechial hemorrhages may be visualized, the initial CT scan often is normal.146 Most of the clinical and experimental studies support that the diffuse injury occurs at the time of initial impact and is not the result of other adverse factors, such as decreased brain oxygenation, increased ICP, or brain swelling,145,147 although these factors tend to accentuate the amount of damage.

The term gliding contusions originally was introduced by Lindenberg and Freytag148 to describe hemorrhagic lesions in parasagittal white matter; this brain injury concept has been reanalyzed.149 Gliding contusions frequently are found in association with DAI and acute subdural hematomas. Two mechanisms have been considered in relation to the formation of these contusions. First, during angular acceleration, more movement and displacement of the cortical gray matter occur than of the deep white matter; accordingly, most of the tissue injury occurs at the gray-white junction because shearing strains predominate. The second mechanism involves excessive displacement of the bridging veins during brain acceleration. Patients with severe DAI who survive may be profoundly disabled, but some patients with mild or moderate DAI may recover with mild or no disability.81 The importance of recognizing DAI in the initial evaluation cannot be overstressed; this will affect the management and outcome. Clinically, these patients will have impaired consciousness secondary to bicortical damage, possibly in association with unremarkable imaging findings and normal ICP.

Penetrating Brain Injury

It is estimated that between 6000 to 7000 people die each year in the United States from gunshot wounds to the brain.130-139,150152


Low-velocity penetrating brain injuries include wounds from nail guns, arrows, and knives and some types of gunshot wounds (Fig. 66.15). Gunshot wounds are divided into low-velocity injuries, such as from civilian handguns and shrapnel, and high-velocity injuries, from rifles and military weapons152 (see Table 66.1).

An Israeli report described an intermediate type of injury from spherical bolts whereby the ball bearings used by suicide bombers, having unique ballistics, may cause a type of “stab wound” injury to the brain.152


The biomechanics of penetrating brain injury (PBI) are important and require an understanding of the dynamics of projectiles—that is, ballistics. Ballistics can be divided into three phases. The first phase, internal ballistics, deals with the source of the projectile and its intrinsic dynamics (e.g., rifle, bomb). External ballistics, the second phase, focuses on the flight of the projectile itself and the deviation of its longitudinal axis relative to the line of flight, referred to as yaw motion. Collision of the projectile with any object during this external phase will change the speed, angle, and yaw. The third phase is referred to as terminal ballistics, which is the physical interaction between the projectile and the body.152

As the projectile directly crushes tissue, it forms a permanent cavity, and as surrounding tissue is compressed, a temporary cavity also is formed. As the missile breaks through the skull, secondary projectiles of bone fragments may be generated, causing independent, secondary permanent and temporary cavities.152-156

Terminal ballistics are influenced by many factors, such as size, shape, and stability of the penetrating object; however, most authors (but not all) believe that entrance velocity is the most important factor in determining the degree of tissue damage.152,153,155,157159 Experimental evidence has demonstrated an immediate increase in ICP with a pressure wave transmitted throughout the cranial cavity, which probably accounts for the transient respiratory arrest observed in PBI.152,160 High-velocity injuries (velocity greater than 320 meters per second) cause shock waves that emanate from the front of the missile. These shock waves can reflect off the inside of the skull and summate, producing significant pressure gradients with remote effects.

Surgical Management

Most of what is known about the surgical management of PBI comes from the battlefield, beginning with World War I, when Harvey Cushing, the father of American neurosurgery, described his technique of en bloc craniectomy under aseptic conditions; thorough debridement of scalp, bone, brain, metal, and bone fragments; and watertight closure of the scalp. Using this management strategy a century ago, in the absence of antibiotics, Cushing reported a significant drop in postoperative mortality from 55% to 28%.161,162

Mortality decreased even further in World War II, primarily because neurosurgical personnel were present in forward military hospitals and antibiotics were introduced. The operative approach, however, was largely the same: that of radical debridement. Treatment consisted of four tiers: emergent life-saving maneuvers through hemostasis and cerebral decompression; prevention of infection through extensive debridement; preservation of nervous tissue through prevention of meningocerebral scars; and restoration of anatomic structures through accurate closure of the dura and scalp. Although untested against other management strategies, this approach became the standard of treatment for PBI by default. Indeed, in the U.S. military, thorough debridement of intracranial bone and metal fragments was official military policy through the Vietnam War.162

The overall mortality rate from PBI sustained during relatively modern wartime conditions ranges from 8% to 43% and generally is considered to be in the 20% range.162-171 The mortality rate from military gunshot wounds has been reported to be 2.5 to 4 times more than from shrapnel.162165,170 Of interest, mortality rates from wartime PBIs in the U.S. military from World War II through the Vietnam War did not change significantly,172 which is consistent with the relatively unchanged management strategy of complete removal of intracranial bone or metal fragments, using repeated surgery as necessary.173

Aggressive debridement was intended to reduce complications such as infection, epilepsy, and cerebral edema. However, studies reveal that additional surgery to remove retained fragments results in significant morbidity and mortality.170,174,175

The use of broad-spectrum antibiotics in recent wars has resulted in data suggesting that retained bone fragments are not independently associated with an increased risk of infection. Therefore, aggressive debridement is no longer supported.162,176 The use of broad-spectrum antibiotics in PBI has now become universal.152

Multiple studies have looked at post-PBI epilepsy and reported an incidence of 22% to 53%.167,177,178 Several studies suggest that it is not necessary to remove all bone fragments in order to decrease the chance of developing epilepsy, but the value of removing metal fragments in this regard remains unclear.162,167,169,177,178 In military PBI, it appears that vigorous debridement is associated with increased morbidity and mortality and is not necessary to prevent infection, has no obvious efficacy in preventing epilepsy, and does not appear to improve survival.162

A CSF leak resulting from a PBI is the variable most highly correlated with intracranial infection.* One study found that most CSF leaks appeared within 2 weeks of the injury and less than half of them closed spontaneously. The incidence of infection was approximately 10 times higher in the group with CSF fistulas than in the group without leaks, and the mortality was greater as well.181 Because CSF leaks are the primary predictor of the development of intracranial infection,167,169,171 it is important to fix or prevent a CSF leak; therefore, a watertight closure of the scalp entry wound is necessary.181

To accomplish this objective, various surgical interventions are available. Small entrance wounds without underlying surgical pathology may be cleaned up and closed in the emergency department.182 Complex entry wounds or those with underlying surgical pathology will require more extensive surgical manipulation, which may include extending the incision to allow vigorous superficial debridement; complete excision of devitalized tissues including muscle, fascia, and periosteum; extension and debridement of the bony entry or exit site by craniectomy; and similarly generous debridement of the dural opening with watertight closure, using grafting as necessary.