Clinical manifestations

Presentation and course

Epidural hematomas arise as a consequence of head trauma in nearly all cases. However, the severity of that trauma is not always a reliable indicator of epidural hematoma formation; minor, even trivial, trauma may be responsible. Although signs of external injury are frequently present and often indicate the site of impact, their absence should not preclude further diagnostic studies. Rarely, nontraumatic epidural hematoma can occur postoperatively as a result of overdrainage of spinal fluid during surgery, particularly in the pediatric and younger adult population (20; 67).

The clinical variability associated with epidural hemorrhage is remarkable. Rarely, epidural hematomas may be asymptomatic. Most present with nonspecific signs and symptoms referable to an intracranial mass lesion. The mode of presentation may be correlated with the size and site of the hematoma, the rate of expansion, and the presence of associated intradural pathology. Epidural hematomas involving the temporal lobe may cause a more precipitous decline than hematomas at other sites, due to their proximity to the brainstem. Finally, concomitant intradural lesions tend to induce elevated intracranial pressure and, thus, create additional neurologic morbidity (41).

Alteration in consciousness is a hallmark of epidural hematoma, but can be variable in extent and duration. Forty percent of patients with epidural hemorrhage present with a Glasgow Coma Score of 14 or 15 (23). The clinical course may follow 1 of five patterns: (1) conscious throughout; (2) unconscious throughout; (3) initially unconscious, then conscious; (4) initially conscious, then unconscious; and (5) initially unconscious, followed by recovery (the so-called "lucid interval"), followed by return to the unconscious state (59; 40; 11). The frequency of each of these scenarios varies significantly among published series. However, the classic "lucid interval" (pattern 5) occurs in less than one-third of patients, and, thus, is not a sensitive diagnostic tool. Furthermore, this lucid interval may occur in conjunction with expanding intracranial mass lesions other than epidural hematoma (11). Increased attention to the risk of significant sustained and cumulative brain injury in high school, college, and professional sports such as football and soccer has led to formation of medical centers that study and treat sports-related cranial impact. The lucid interval following sports injury is well described in a case report by the Vanderbilt Sports Concussion Center, highlighting the delayed deterioration in a 16 year old following his collision with another player’s head during soccer (62).

Clinical manifestations other than impaired level of consciousness may be encountered with epidural hemorrhage, although these are typically late findings. Abnormalities in pupillary function may occur as the lesion evolves, due to impingement of the oculomotor nerve against the tentorial incisura. Anisocoria is noted in 50% to 75% of patients with epidural hematoma prior to operation; in approximately 85% of these, mydriasis is found ipsilateral to the side of hemorrhage (63; 59; 31; 40; 11; 53; 72). Bilateral pupillary dilatation is a relatively unusual and often preterminal event. Lateralized motor deficits are found in roughly one-third of patients (range 25% to 57%); when present, hemiparesis is contralateral to the side of hemorrhage in the vast majority of cases, but rarely may be ipsilateral or bilateral (63; 40; 23; 53; 72). Vegetative motor posturing (decortication, decerebration) is identified in less than one fifth. Disturbances of cardiovascular function and ventilatory drive occur in approximately 15%.

Signs and symptoms referable to epidural hemorrhage generally evolve rapidly after head trauma. Nearly two thirds of patients exhibit prominent findings prompting surgery within six hours of injury; 85% declare themselves within 24 hours (11). The remainder constitute (1) those with small occult lesions that remain asymptomatic, (2) those with slowly accumulating clots that become symptomatic after 24 hours, and (3) those whose hematomas develop in delayed fashion. The third phenomenon is not as rare as once thought among head-injury patients. Although most patients demonstrate radiographic evidence of epidural hemorrhage on their initial CT examination, a small subset develop hematoma collections from 2 to 120 hours (median 12 hours) after presentation, usually after treatment of elevated intracranial pressure by medical or surgical means (71). Thus, intense surveillance is required to detect clinical deterioration in patients with initially unremarkable CT scans obtained soon after injury because such a decline may indicate the delayed accumulation of an epidural hematoma.

Clinical vignette

Vignette 1. A 4-year-old boy was brought to the emergency department within an hour of falling from a height of approximately 4 feet. His mother brought him in because he was initially acting lethargic but was now acting more appropriate. On initial evaluation he was found to be Glasgow Coma Scale 15 (GCS), neurologically intact. There was ecchymosis and apparent small cephalhematoma at the right parietal region. The patient was taken for CT of the head given the nature of the injury and initial lethargy. While still in the emergency department, the patient became drowsier, and over a period of less than five minutes he was unresponsive. He was taken immediately for a head CT, which revealed a greatly expanded mass causing significant midline shift. The patient was taken promptly to the operating room for surgical decompression. Postoperatively he was managed in the pediatric intensive care unit for a few days and subsequently discharged home, neurologically intact.

Vignette 2. A 45-year-old right-handed man presented to the emergency room three days after apparently being assaulted in the park, complaining of progressive headache. He appeared agitated and had biparietal bruises on the scalp and right periorbital ecchymosis. He had no memory of hitting his head or of being beaten by assailants. His neurologic examination was nonfocal other than asymmetric pupils with a larger, slightly sluggish right pupil. A noncontrast head CT showed a large lens-shaped mass arising from the right temporal floor, adjacent to the temporal convexity bone, with right-to-left temporal lobe shift and uncal herniation, causing deformation of the lateral margin of the brainstem. A small, linear, low temporal convexity fracture was seen as were a few punctuate surface hemorrhages in the opposite frontal and temporal lobes.

He was taken immediately to the operating room and underwent a right temporal craniectomy for evacuation of the epidural hematoma. Findings were a small temporal squama fracture with oozing blood through the fracture line and a large, partially congealed clot extending nearly the length of the temporal fossa, causing significant displacement of the dura and temporal lobe. Evacuation of the clot revealed it to be tamponading an actively bleeding, torn middle meningeal artery in the convexity dura, which was cauterized. The patient made an uneventful recovery, and postoperative CT showed complete resolution of the clot and temporal shift.

Biological basis

Etiology and pathogenesis

The overwhelming majority of epidural hematomas arise secondary to direct cranial trauma and typically have an overlying skull fracture. A secondary traumatic source of epidural hematoma can occur after surgical decompression of a contralateral acute subdural hematoma. In this case, the subdural hematoma is evacuated contralateral to the side of a skull fracture, and postoperative CT reveals presence of a newly formed epidural hematoma at the site of the fracture (79). A limited number of cases occur spontaneously, typically the result of hemorrhage from tumors or vascular anomalies involving the skull or dura mater.

Epidural hematomas usually result from direct head injury. In most series from industrialized countries, motor vehicle accidents predominate, followed closely by falls (63; 59; 40; 11; 74; 23; 53; 72). Assaults, sporting accidents, and birth injuries account for the remainder. The mechanism of injury is somewhat age dependent, with falls accounting for a relatively greater proportion of cases in young children and the elderly (40). Although rare, causes of nontraumatic epidural hematoma include: localized infections (causing arteritis), vascular malformations, malignant and metastatic tumors, rheumatologic disorders, sickle cell disease, and patients undergoing hemodialysis or open-heart surgery (08).

Irrespective of the precise nature of inciting trauma, a recognized or occult direct blow to the head is usually essential for epidural hematoma formation. The force required to produce such a lesion is variable; however, it must be sufficient to deform the skull and strip the underlying dura away from the inner calvarial table. Once the dura is separated from the inner table, a space (the "epidural space") is created in which blood may accumulate (30). Classically, the force imparted to the skull exceeds its absorptive and elastic capacities and fracture results. However, about 10% of adults and 20% to 40% of children with epidural hematomas do not have concomitant skull fractures (59; 11). The lower coincidence of fracture and epidural hematoma in children is probably the result of greater skull elasticity, open fontanelles, and unfused sutures in this age group. However, in children the force of head injury, although not always producing a fracture, still can cause the dura to separate from the cranium and create the potential epidural space (16).

As the skull is deformed and the adherent dura forcefully detached, dural-based vessels may be torn or avulsed, with subsequent hemorrhage into the preformed epidural space. The source of bleeding is variable and may be arterial, venous, or likely both. It has always been believed that epidural hematoma in the supratentorial compartment is due to bleeding from the middle meningeal artery. This is probably the case, at least in part, but the concept of isolated arterial bleeding has been called into question. A study examining 29 cadaveric specimens revealed that the “middle meningeal artery is accompanied by a pair of dural sinuses throughout the majority of its course, thus, making exclusively arterial rupture an anatomic improbability” (29). Although disruption of a dural venous sinus is not primarily responsible for supratentorial epidural hematomas, sinus bleeding is implicated in the majority of those arising within the posterior cranial fossa (54). A discrete source of hemorrhage is not always identified at surgery, and occasionally multiple foci of bleeding are present. Simple skull fracture through the vascular diploic space can produce enough epidural mass effect with hematoma to require surgical evacuation, particularly in younger adults (81).

A 2007 study of 24 patients with epidural hematoma, treated nonoperatively, revealed that all patients showed angiographic evidence of vascular injury within the first 19 hours after injury. Seven of the patients had evidence of pseudoaneurysm, and 17 patients had active extravasation of contrast medium (26). It is thought that expansion of epidural hematoma is caused by rebleeding or continued bleeding likely from sources such as these, but the clinical sequelae of this type of finding is not fully known.

The expanding extradural lesion only partially accounts for the neurologic morbidity observed with epidural hematomas. Coincident intradural pathology is encountered in up to 50% of cases and is associated with lower admission Glasgow Coma Score, more substantial and prolonged intracranial pressure elevation, and higher mortality (24; 74; 53; 72). Intradural lesions often identified concomitantly with epidural hemorrhage include subdural hematoma, parenchymal contusion, and diffuse axonal injury (11; 72). In general, it is the sequelae of these lesions that dictate the degree of residual functional impairment in patients who survive epidural hematoma.

Finally, there is the relatively rare occurrence of double acute epidural hematoma—at least two separate epidural hematomas occurring at once in either unilateral or bilateral fashion. In a 2004 retrospective review of 1,025 epidural hematomas, there were 46 (4.48%) double epidural hematomas. The mortality for double epidural hematoma was 34.8% versus only 9% in single epidural hematoma. The majority of the double epidural hematomas were bilateral and required surgery on both sides, and it was noted that lateralization of neurologic exam was less common than in single epidural hematoma (39).

A rare cause of intracranial epidural hematoma in which no trauma is recognized is acute or chronic leakage of a pseudoaneurysm of the middle meningeal artery in which a prior trauma to the artery likely occurred (83).

Epidemiology

The incidence of epidural hematoma following head trauma varies with the population surveyed. When all patients gaining hospital admittance for head injury are considered, epidural hematoma makes up only 1% to 3% of cases (40; 24). If only comatose (Glasgow Coma Score less than or equal to 8) patients are included, this incidence rises to nearly 10% (74).

The age distribution of patients with epidural hematoma reflects that of head injury in general. It is a disease of the young adult, especially those between the ages of 10 and 40 years (59; 40; 24; 11). Epidural hemorrhage is less frequent in children younger than two years of age and in adults older than 60 years of age, owing to the intimate adherence of the dura mater to the inner calvarial table at both extremes of life (31; 16). The relative elasticity of the infant skull may contribute to the rarity of epidural hematoma in this group as well. Males are affected far more frequently than females, with an overall gender ratio of approximately 4:1 in most series (59; 40). The gender predilection is much less marked at the extremes of age, approaching a 1:1 ratio (59). Of note is that as the elderly population increases, we are likely to see an increase in all disease states in the elderly, and epidural hematoma is no exception. A 20-year retrospective study out of Finland looking at nearly 8000 craniotomies for trauma in all patients 18 years and older found that the median age for acute traumatic subdural hematoma was 67, versus 45 for acute traumatic epidural hematoma. Acute epidural hematoma represented 14% of this large surgical series, with incidence in men exceeding women by greater than 3 to 1. This study also showed only 1.2% of craniotomies for acute epidural hematoma in ages 70 or older (66). However, another large retrospective institutional review of 3249 epidural hematomas also revealed that only 32 (less than 1%) were in patients over the age of 65 years. In this group, assault was the most common inciting factor, and no patient in a coma or over 75 years of age had a good outcome (51). Motor vehicle accidents, or “road brain trauma,” has a particularly high incidence of epidural hematoma, often associated with underlying brain injury such that simple evacuation of epidural hematoma, as seen after most lower impact sports injuries, will predictably have worse outcomes overall. A 10-year Saudi Arabian study of 520 craniotomies following road accidents found that 42% of patients with traumatic brain injury had extra-axial hematoma; the majority were for acute subdural hematoma, and the best prognosis was following evacuation for epidural hematoma (01).

Epidural hematomas may occur at any site in the cranial vault. The temporal region is involved in roughly 70% of cases, with the hematoma usually accumulating under a fractured squamous temporal bone (59; 40). Temporal tip epidural hematomas have been shown to be associated with a greater incidence of zygomatic arch fractures, lateral orbital cavity fractures, skull-base fractures, and cranial nerve injury (87). Additionally, the squamosal portion of the temporal bone is relatively thin, likely sustaining fracture with a lesser force than other areas of the cranium. Approximately 15% occur over the frontal convexity, 10% are parieto-occipital, and the remainder is parasagittal or infratentorial.

Lesions within the posterior fossa account for a small portion of all epidural hemorrhages, are almost always associated with occipital skull fractures, and usually arise from disrupted dural venous sinuses and fracture lines (65; 56). In the pediatric population, axial CT scanning alone may lead the clinician to confuse posterior fossa epidural hematoma with venous sinus thrombosis, as the hemorrhage may peel the dural sinus away from the inner table of the suboccipital bone, compressing but not occluding the sinus. CT venography or MRV will assist in making this critical distinction as anticoagulation for venous sinus thrombosis may prove devastating if the actual diagnosis is epidural hemorrhage (77). Occipital skull fracture has been found in 85% to 95% of posterior fossa epidural hematomas (56; 42). They can often be clinically silent, producing only generalized complaints. Multiple single institution studies of all patients with epidural hematomas found that between 2.7% and 9.8% of patients had a posterior fossa epidural hematoma. In a group of 65 posterior fossa epidural hematomas, 53 patients were treated surgically and 12 conservatively; four of the patients initially treated conservatively decompensated and required surgery (42). The decision to pursue conservative therapy was generally based on patients having no sign or symptom of brainstem compression, open basal cisterns, and no hydrocephalus. Hydrocephalus has been found to be an ominous sign, and children tend to have better outcomes than adults (56). Usually, epidural hematomas are unilateral; however, bilateral lesions may be identified in up to 5% (33).

Prevention

Preventive strategies for epidural hemorrhage and for head injury in general are inextricable, as one does not occur without the other. Given the often devastating and irreversible nature of the primary traumatic injury (created by the initial impact), the most effective means to mitigate brain injury is through prevention. Most states have passed legislation mandating seat belt use in automobiles and helmet use for motorcyclists. Infant restraint seats are now required throughout the United States, and most new cars are equipped with air bags. Helmet use has increased dramatically among bicyclists, in both the child and adult populations. Most head trauma outcome studies that compare use or nonuse of helmets during sports, such as cycling, soccer, or horseback riding, show at least a trend toward fewer instances of loss of consciousness or epidural hematoma (05). Looking specifically at horseback riders—a sport slow to adopt helmet use-- a British group studied 40 patients with head injury and noted a significant increase in number and extent of skull fractures in the non-helmeted group and a 5-fold incidence of intracranial hemorrhage in those without helmet protection (07). Citing an annual incidence in the United States of 200,000 pediatric cycle accidents resulting in 22,000 traumatic brain injuries, one group looked at the impact of 10 years of helmet legislation in the state of Illinois (1999 to 2009); helmet use was associated with a lower incidence of traumatic brain injury in the group of 3080 children treated for bicycle crashes during this period. However, only 5% of these children reported use of a bicycle helmet, indicating an obvious problem with compliance and enforcement of the helmet law, and there was a significantly lower incidence of reported helmet use among children of lower socioeconomic status. In fact, legislation was shown to have no impact on helmet use (86). Finally, prevention of alcohol-related traffic fatalities through drunk-driving laws and more vigorous enforcement policies will likely lead to a lower incidence of epidural hematoma. These efforts have made a positive impact on the head injury epidemic in the United States; however, further progress must be made in these and other preemptive endeavors (44).

Differential diagnosis

Because of the variability in presentation, essentially all traumatic intracranial lesions must be included as potential diagnostic considerations. Principal among these are acute subdural hematoma, parenchymal contusion, intraparenchymal hemorrhage, and diffuse axonal injury. The difficulty in distinguishing these entities is compounded further by the fact that they often coexist in the same patient (63; 24; 74; 53; 72).

Occasionally, a clear history of trauma is not forthcoming; in such instances, spontaneous subarachnoid hemorrhage, hypertensive or angiopathic intraparenchymal hemorrhage, infarct, seizure, infection, and cerebral neoplasia should also be considered. All may present with depressed consciousness, headache, and various degrees of focal or global neurologic dysfunction, and, thus, may mimic epidural hematoma clinically. The additive psychotropic effects of alcohol or drugs may further confound the assessment of mental status. Systemic derangements that may alter consciousness such as hypotensive or hypovolemic shock, profound hypoglycemia, or diabetic ketoacidosis also should be considered in the appropriate context.

Diagnostic workup

The presence of an acute epidural hematoma should be suspected in any patient presenting with external evidence or history of head trauma, particularly when consciousness is impaired or a focal neurologic deficit is present.

CT imaging is currently the most effective means of imaging the intracranial compartment in the acute trauma setting. Skull fractures also may be identified by CT and are best demonstrated with “bone window” attenuation settings.

Intravenous contrast administration is not recommended as enhancement may obscure associated subarachnoid and parenchymal hemorrhages. Characteristically, epidural hematomas appear as hyperdense, lentiform (biconvex), extra-axial masses, with margins limited by the skull sutures, and located within the temporal region. Compression of the underlying brain is characteristic; often, the ipsilateral ventricular system is effaced and midline structures are shifted toward the contralateral side. When present, coexistent intradural pathology is identified.

Indications for emergent cranial CT imaging in the initial evaluation of the head-injured patient include the following: (1) history of loss of consciousness; (2) depressed level of consciousness; (3) focal neurologic deficit; (4) deteriorating neurologic status; (5) skull fracture noted on exam or obvious penetrating injury; (6) worsening or severe headache; (7) persistent nausea or vomiting; (8) posttraumatic seizure; (9) mechanism of injury suggesting high risk of intracranial hemorrhage; (10) examination obscured by alcohol, drugs, metabolic derangement, or postictal state; (11) patient inaccessibility for serial neurologic examinations; (12) coagulopathy and other high risk medical conditions, and (13) patient is an otherwise poor or elusive historian. CT evaluation should proceed as soon as the patient is hemodynamically stable and immediately life-threatening injuries have been addressed. As the incidence of delayed epidural hematoma formation following head trauma is substantial, any deterioration in the neurologic examination warrants prompt evaluation by CT, even if a previous study was normal. Delayed onset of epidural hematoma has been reported in patients whose initial CT scans were normal (09). Although unusual, a CT scan may on occasion indicate acute bleeding by the so-called “swirl sign,” which is described as a hyperacute volume of blood density mixed with a volume of slightly lower density of clotted blood and is another indicator for urgent surgical intervention (34) because of poor prognosis without surgery (85).

CT angiography (CTA) has been recommended as an early or first diagnostic study in the acutely head-injured patient, rather than the traditional non-contrast head CT, because of the additional information given by the contrast. Extravasation of contrast into the hematoma or into the surrounding brain has been associated with a poorer prognosis, as well as with a higher likelihood of subsequent hematoma expansion, whether extra-axial (epidural, subdural) or intraparenchymal (52; 73).

The role of MR imaging in the evaluation of acute head trauma is nominal. Presently, MR imaging is best suited for defining associated parenchymal injuries, such as diffuse axonal injury, in the days following the acute event; however, MR even for these indications rarely changes management or outcome. Current techniques in MRI allow higher resolution sequences, such as diffusion tensor imaging that now allows detection of deep white matter injuries, and more trauma-specific local changes in tissue chemistry can be seen with MR spectroscopy (49).

Other, more traditional, diagnostic tools have largely been supplanted by cranial CT in the initial assessment of the head-injured patient. Plain skull radiographs are inexpensive and easily obtained, and often demonstrate fractures in patients with epidural hemorrhage. However, the predictive value of such films is poor, as one finding is not requisite for the other. Lumbar puncture, once considered an indispensable adjunct in the evaluation of the head-injured patient, is now relatively contraindicated in this setting, owing to the risk of precipitating transtentorial herniation with the procedure. Angiography is not used as part of the standard evaluation for epidural hematoma.

Application of a new technology, “infrascanner,” has been studied in a multinational trial and is in use worldwide by U.S. Marine battalion aid stations to detect early intracranial hemorrhage following trauma (03). The technology is based on the ability of near-infrared light to transcranially detect surface changes in the brain, such that initial results indicate very high sensitivity and specificity for detecting brain surface hematomas in adults and children.

Management

There are no prospective, randomized trials regarding surgical management of epidural hematoma, and all management strategies should be tempered with this understanding. The initial management of the patient with epidural hematoma depends on the following factors: (1) the clinical condition of the patient, (2) the size and nature of the hematoma, and (3) the time since head injury. Immediate surgical evacuation is usually indicated in the patient in whom neurologic abnormalities are thought to be the result of an epidural hematoma, especially if his or her status is worsening. It is generally believed that large hematomas, even if asymptomatic, should be drained urgently; continued bleeding can cause rapid deterioration. The presence of unclotted blood on early CT scans suggests continued bleeding and the need for immediate surgery (37; 64). Evacuation is more likely to be needed for epidural hematomas within hours of trauma than for those several days old. In the comatose patient, standard acute management of severe head injury should be instituted concurrently with preparations for surgery.

The use of prophylactic antiepileptic medication in the context of traumatic brain injury, and for the subset of traumatic epidural hematoma patients, is still controversial. A prospective comparison of 522 patients with “blunt” (nonpenetrating) brain injury--272 in the United States and 250 in China treated and not treated prophylactically, respectively—showed no difference in the small number of patients who developed seizures (3.7% vs. 2.8%), with a slightly higher incidence of seizures correlating with severity of injury measures such as Glascow Coma score rather than with use of prophylactic antiseizure medication (48).

Surgical evacuation of epidural hematoma is usually by way of a craniotomy, centered over the point of maximum clot thickness. Craniectomy, using bone rongeurs to expand a burr hole, was once a common procedure (45). The use of reconstructive titanium mesh allows bony reconstruction at the completion of surgery; alternatively, the use of power instruments offers intact bone flap removal and avoids leaving a skull defect. A particularly rapid approach to epidural hematoma evacuation involves craniotomy via linear incision (27). Some of the clot adheres to the elevated bone flap, and the remainder of the hematoma is evacuated with suction or irrigation or a cup forceps. Exposure must be ample to control the bleeding source and allow total clot removal without brain retraction. If preoperative imaging studies suggest a concomitant subdural hematoma or resectable parenchymal contusion, or if the dura remains tense or exhibits a bluish hue, subdural exploration is indicated. With advances in neurosurgical endoscopy, there has been an interest in evacuation of acute traumatic epidural hematoma through simple burr hole; the surgeon must anticipate the need to control active arterial bleeding, which often occurs at the moment of clot decompression (43; 82).

Posterior fossa epidural hematomas may be accessed via suboccipital craniectomy, using either a paramedian (if the hematoma is unilateral) or midline (if bilateral) incision.

As these lesions tend to extend both above and below the tentorium, preparation must be made to cross the transverse sinus with the exposure. The initial burr hole is placed over the point of maximal clot thickness, but should not lie directly over a dural venous sinus. This trephine is expanded as above, with bone over the involved sinus removed last, and the underlying hematoma evacuated. Intradural pathology may then be addressed if necessary. Obstructive hydrocephalus, a relatively prevalent complication of posterior fossa epidural hemorrhage, may necessitate CSF diversion via ventriculostomy, although hematoma evacuation typically is sufficient to re-establish normal CSF flow.

Burr hole with prolonged epidural drainage has been advocated in patients with coagulopathy and history of acute trauma, with CT finding of isodense epidural hematoma. One report cited favorable outcome for eight coagulopathic patients treated acutely with three days of continuous epidural drainage as an alternative to open craniectomy or craniotomy evacuation of clot, taking advantage of the nonclotting nature of the epidural blood in these patients (36). However, rapid reversal of the coagulopathy with open surgical evacuation would seem preferable in most cases.

Postoperative care is dictated by the extent of intradural injury and degree of residual neurologic impairment. If preoperative imaging suggests the presence of parenchymal contusion, hemispheric edema, or diffuse axonal injury, an intracranial pressure monitor or ventriculostomy is placed at the time of surgery to facilitate postoperative intracranial pressure management. As an adjunct, hemicraniectomy is sometimes considered in the trauma population, and further studies to delineate its usefulness are presently ongoing. Aggressive medical treatment of persistent intracranial hypertension, including osmotic diuresis, blood pressure support, sedation, and neuromuscular blockade, is critical to optimize outcome (58; 60). Any postoperative neurologic deterioration or failure to recover function as anticipated mandates repeat neuroimaging.

The management strategies presented are applicable to the vast majority of patients with traumatic epidural hematomas. Under rare circumstances, however, it may be necessary to deviate from these options. Fortunately, there are few clinical scenarios in which a patient's condition is deteriorating so rapidly as to render diagnostic studies unobtainable. If confronted with such a situation in which transtentorial herniation and death are imminent, immediate exploratory trephination is indicated as a life-saving measure. The initial burr hole is placed in the temporal fossa just above the zygomatic arch on the side ipsilateral to the dilated pupil and contralateral to the hemiparesis. If both signs are present on the same side, trephination should be performed ipsilateral to the dilated pupil, as this is a more reliable localizing sign (46). If neither sign is present, the burr hole is placed on the side of skull fracture. Subsequent burr holes are placed in the frontal and parietal regions in such a manner as to facilitate incorporation into a formal craniotomy if hematoma is encountered, as burr hole exploration is a temporizing measure only. Intraoperative ultrasonography may improve the diagnostic accuracy of this procedure (02). If no hematoma is encountered, trephination at similar sites should be performed on the contralateral side; bilateral suboccipital burr holes are indicated when no supratentorial hematoma is revealed. A CT scan is mandatory if the exploratory burr holes fail to reveal epi- or subdural hematoma.

Solutions for emergency treatment centers lacking prompt neurosurgical care are emerging, including the use of simple devices for penetrating bone over the epidural collection and drainage system to evacuate the accumulated blood. One such system was used successfully with virtual neurosurgical guidance using videoconferencing (32).

Given that there are no prospective, randomized trials regarding surgery, only recommendations are available. Recommendations representing members of the American Association of Neurological Surgeons, Congress of Neurological Surgeons, the European Brain Injury Consortium, American College of Surgeons (Committee of Trauma), and the World Federation of Neurological Surgeons are: evacuation of an epidural hematoma greater than 30 cm3, regardless of GCS; nonoperative management with close neurologic observation in a neurosurgical center if a patient has an epidural hematoma with GCS greater than eight and no focal deficit and less than 30 cm3 hematoma and less than 15 mm hematoma thickness and less than 5 mm shift (12).

Several investigators have reported successful resolution of epidural hematomas with nonsurgical treatment (14; 37; 19; 75; 47). Not all epidural hematomas require surgical decompression, often not just one factor dictates management, and in some situations it is appropriate to pursue conservative management. Generally, this approach is advocated only for those patients with small hematomas in extratemporal locations who exhibit no or minimal clinical findings and stable or improving neurologic conditions or for those with associated trauma and other conditions, in whom extended hospital observation will be necessary in any event. In Brazil, 80 consecutive patients were treated successfully without craniotomy by performing endovascular embolization of the extravasating middle meningeal artery, such that follow-up consistently demonstrated no increase in size of the hematoma following embolization; this is in contrast to Peres and colleagues’ review of the literature, in which 82 of 471 patients (17.4%) observed with untreated epidural hematoma came to surgical evacuation (70). Interventional middle meningeal artery embolization is still considered for controlling a slowly expanding hematoma and is thought to be more of a “salvage” procedure to avoid open surgical evacuation in the elderly or higher risk surgical patients (68).

A similar approach has been successfully employed in children with epidural hematoma (06). A study of 13 children with epidural hematomas, all greater than 1.0 cm, who were conservatively managed (1 converted to surgery for worsening headache, deteriorating vigilance, and increasing size of hematoma on CT) showed that patients and parents were completely satisfied with the outcome, there were no head injury-related complaints, and there was no epilepsy (04). The basis for such an approach is that, for reasons already mentioned, children tolerate increases in intracranial pressure better than adults, and size alone should not be the only consideration when deciding between operative or nonoperative management. In a study of 31 infants (younger than 2 months of age) with epidural hematoma, seven were managed nonoperatively (21). The criteria for conservative management include normal neurologic examination and size of the hematoma not meeting radiographic criteria for decompression. There is still debate regarding the size of epidural hematoma in children that can be managed conservatively. One study out of the University of Montreal found that of 16 pediatric patients with epidural hematoma greater than 15 mm thickness, only two patients required surgery after a period of observation to be managed successfully (17). Additional criteria should include no or only mild signs of elevated intracranial pressure, and patients should be observed in an intensive care unit with neurosurgery readily available (04). A level 1 trauma center looked retrospectively at 195 pediatric patients with epidural hematoma admitted for observation, of which 37 progressed and needed surgical evacuation, and the remaining 158 were managed well conservatively. The large majority of patients who did not require surgery were neurologically intact with epidural hematoma less than 15 ml volume and without mass effect (15).

As a rare but significant complication of conservatively managed epidural hematoma in children, ossification of nonabsorbed epidural hematoma may occur, requiring delayed surgical intervention (22). Symptomatic mass effect from the ossified lens-shaped ossification is likely to require surgery: an Italian study reviewed 16 studies from the available literature of epidural hematoma series and found 18 pediatric cases, of which all but when required surgical resection. The recommendation was for MRI or ultrasound at two months following initial diagnosis and conservative management of an epidural hematoma, to avoid missing this early complication (25). A University of Texas level 1 trauma center looked retrospectively at 117 pediatric patients presenting with a Glasgow Coma Scale of 14 or better and found that 24 were taken to surgery and the remaining 93 were managed conservatively (10). The significant factors favoring surgical treatment were decline in GCS (40%), hematoma size alone (21%), localizing neurologic deficit (20%), progression on repeat imaging (12%), and other (7%). Looking at hematoma thickness alone, greater than 95% of patients with thin (0.15 biparietal skull diameter or less) epidural hematomas were treated without surgery. A rare form of epidural hematoma occurs in children subjected to extreme flexion-extension injuries in which the tectorial membrane is stripped from the clivus, forming a retroclival hematoma; the diagnosis may be missed on CT as an expanded density behind the clivus, better visualized by skull base or cervical MRI. Although these are successfully managed without surgery, the diagnosis is critical as anticoagulation is to be avoided (28).

The common use of repeated head CT in patients under observation for head injury has a very low yield. A 2017 Harvard study, which included 380 mild traumatic brain injuries with first CT scan “positive” but managed nonoperatively, showed that only three patients went on to surgery after CT showed significant worsening with expanding hematoma, and each of these patients also showed significant neurologic decline. Their recommendation was to observe this group of head injured patients for an additional eight hours and only obtain follow-up CT for those patients who demonstrate neurologic decline (Stippler et 2017). Further corroborating studies are needed.

Surgery, with proven effectiveness and safety, remains the gold standard (11).

Venous epidural hematoma originating from a torn venous sinus, generally produced by an overlying fracture, presents a particular challenge because the expanding hematoma may lead to sinus compression and occlusion, even if the mass effect on the underlying brain does not warrant surgical decompression. One large series of 403 patients with transsinus fracture studied with either CTV (CT venography) or MRV (MR venography) found patients to have a significant mortality in the delayed diagnosis group (diagnosis and treatment on average 7 days after trauma). Nonoperative patients were given intravenous thrombolysis with low-dose urokinase with imaging resolution of sinus thrombosis in the early diagnosis group (diagnosis and treatment within 3 days of trauma), which correlated with favorable clinical response (84). A retrospective study of 268 cases of traumatic epidural hematoma identified 32 cases (12%) of venous origin (23 supratentorial and 9 infratentorial); 10 cases were managed with surgical evacuation and 22 without. The decision for surgery was based on size of hematoma, location of hematoma and evidence of expansion on serial imaging, impaired neurologic status or evidence of decline; overall outcomes were excellent in both groups using these criteria for management (69). A Taiwan neurosurgical group has addressed the challenges of vertex epidural hematoma produced by a torn sagittal sinus, and of a combined supratentorial-infratentorial epidural hematoma produced by a torn transverse sinus, by performing a “strip craniotomy” in which a strip of bone is deliberately retained over the sinus to allow compressive suturing of the dural sinus to the overlying bone (80).

Outcomes

The natural history of large traumatic epidural hematomas is dismal if the lesion is unrecognized or untreated. In the vast majority of cases, progressive neurologic dysfunction is precipitated by the expanding mass lesion, ultimately resulting in transtentorial or uncal herniation, brainstem compression, and death. Rapid diagnosis and prompt surgical evacuation afford the best chance for optimizing outcome. This is particularly crucial for those patients with minimal intradural pathology whose neurologic dysfunction may be attributed solely to the expanding extra-axial lesion. If treatment is instituted prior to obtundation, pupillary dysfunction, or vegetative motor posturing, the probability of full functional recovery is high (44).

In series amassed since the advent of cranial CT, favorable outcome following operative intervention for epidural hematoma has been reported in 55% to 94% of cases (11; 74; 23; 53; 72; 61; 76). Mortality has ranged from 5% to 41%. Factors that appear to have the greatest negative influence on outcome include (1) preoperative coma (Glasgow Coma Score less than or equal to 8), (2) preoperative pupillary dilatation (bilateral greater than unilateral), (3) severe preoperative motor deficit or vegetative posturing, (4) age greater than 40 years, (5) presence of coexistent intradural pathology, (6) elevated intracranial pressure, (7) short latent period between injury and first symptoms, and (8) large hematoma (greater than 150 mL). Small epidural hematomas, not large enough to cause significant cerebral compression or symptoms, often resolve spontaneously (37; 19), which is likely to occur within 2 or 3 months (04). Mortality of children with epidural hematoma is much lower than that of adults, and outcomes are significantly better (61). The exception for children is the posterior fossa hematoma, which can be devastating even at smaller volume; therefore, early CT scanning in the acute setting of nausea and vomiting and impaired mentation is still advised; early evacuation of posterior fossa epidural hematoma in children has favorable outcome (18). A further complicating feature of very young children requiring surgical evacuation of epidural hematoma is underlying brain ischemia from hematoma mass effect, leading to infarct after surgery. A single-center retrospective study from Paris over a decade of head trauma found 48 infants less than 18 months old who required surgical evacuation of epidural hematoma; of those, 17 (36%) had MRI evidence of cortical ischemia due to hematoma pressure effect, a predictor of residual neurologic sequelae (50). Hyperperfusion with brain swelling and parenchymal hematoma formation has also been described after hematoma evacuation in both older adults and young children. One series reported from China found that 12 of 157 patients developed hyperperfusion syndrome within six hours of epidural hematoma evacuation, some leading to herniation if preemptive measures were not taken, such as steroids and hyperventilation (38).

Special considerations

Pregnancy

No data that address epidural hemorrhage in the pregnant patient have been presented in the literature to date. Due to the lethality of the condition, however, it is imperative that the diagnosis and treatment of epidural hemorrhage not be delayed for fear of harming the fetus, as maternal death (and, thus, fetal demise) may result.

Anesthesia

The anesthesiologist's role in the management of epidural hemorrhage is critical. Endotracheal intubation must be accomplished with minimal stimulation to avoid marked intracranial pressure elevation. Short-term hyperventilation to a PaCO2 of 25 to 30 torr may reduce intracranial pressure in patients with a rapid clinical deterioration. In general, hyperventilation is no longer recommended, with a target PaCO2 in the normal range (13; 57). A smooth anesthetic induction without hemodynamic instability optimizes cerebral perfusion pressure. Generally, relaxation is provided by a nondepolarizing neuromuscular blocking agent. Volatile anesthetic agents should be avoided due to their vasodilatory effect on cerebral blood vessels. Judicious fluid management is crucial to maintain cerebral perfusion pressure without contributing to cerebral edema. This may be particularly challenging when mannitol has been given to promote osmotic diuresis. Finally, the anesthetic should be titrated so that the patient's neurologic status can be assessed immediately following completion of the surgical procedure.