Blunt cerebrovascular injury: Mechanisms, screening, and diagnostic evaluation

INTRODUCTION — Blunt carotid and vertebral artery injuries, collectively termed blunt cerebrovascular injury, are rare but potentially devastating events. In the era prior to screening, blunt carotid injury was associated with mortality rates ranging from 23 to 28 percent, with 48 to 58 percent of survivors suffering permanent severe neurologic deficits [1].

The overall incidence is low in patients sustaining blunt trauma. Clinical studies in the early 1990s suggested that these injuries were being underdiagnosed [2]. Increased recognition through diagnostic screening based upon specific clinical criteria has increased the reported incidence to approximately 1 to 3 percent in patients with blunt trauma admitted to trauma centers [1,3-5].

Injury mechanisms, screening, and diagnosis of blunt cerebrovascular injury will be reviewed here. The treatment of blunt cerebrovascular injury is discussed separately. (See "Blunt cerebrovascular injury: Treatment and outcomes".)

The diagnosis and management of penetrating cerebrovascular injury and spontaneous cerebrovascular dissection are reviewed elsewhere. (See "Penetrating neck injuries: Initial evaluation and management" and "Cerebral and cervical artery dissection: Clinical features and diagnosis".)

CEREBROVASCULAR ANATOMY — The vascular supply to the brain is divided into the anterior and posterior circulations originating from the carotid and vertebral arteries, respectively. The circle of Willis connects the anterior and posterior circulations but is completely intact and symmetric in only approximately 20 percent of individuals [6]. The anatomic variability of the collateral circulation may help explain the clinical presentations of patients with cerebrovascular injuries and underscores the need for complete imaging of cerebral circulation when injury is suspected [7]. (See 'Imaging evaluation' below.)

Anterior circulation — The anterior circulation supplies the majority of the cerebral hemispheres except the occipital and medial temporal lobes. Injury to the vessels of the anterior circulation can lead to ischemic or hemorrhagic hemispheric stroke. (See 'Clinical presentation' below and "Stroke: Etiology, classification, and epidemiology".)

The left common carotid artery (CCA) originates from the aortic arch, whereas the right CCA originates from the brachiocephalic trunk (innominate artery) (figure 1). The CCA divides into the internal carotid artery (ICA) and external carotid artery (ECA) at the level of the superior border of the thyroid cartilage corresponding to the disc space between the third and fourth cervical vertebral bodies (C3/C4).

The external carotid artery (ECA) has multiple branches that supply the face and scalp (figure 2) and provide collateral circulation to the brain. Traumatic injuries to the ECA are usually tolerated neurologically with the occasional exception of patients with preexisting cerebrovascular disease. When the ICAs or vertebral arteries (VAs) have significant stenoses or occlusion, the ECA branches may provide critical collateral pathways for cerebral blood flow (figure 3).

Internal carotid artery segments — The ICA (figure 4) is divided into four segments based upon anatomic landmarks.

●Cervical – The cervical portion usually has no named branches, as it ascends anterior to the spine. Blunt trauma to the neck in this region can lead to compression of the ICA against the transverse processes of the first through third (C1 to C3) vertebral bodies.

●Petrous – The petrous segment traverses the carotid canal in the petrous portion of the temporal bone. Basilar skull fractures can cause laceration of the ICA at this site.

●Cavernous – The cavernous portion, also called the carotid siphon because of its S shape, is the first portion of the ICA within the cranial vault. This segment is suspended between the layers of the dura mater that form the cavernous sinus. Fracture of the sphenoid bone may injure the cavernous portion.

●Cerebral – The ICA perforates the dura mater at the anterior clinoid process to become the cerebral (or supraclinoid) segment, which divides into the anterior and middle cerebral arteries [8].

Posterior circulation — The posterior circulation supplies the midbrain, cerebellum, occipital lobe, and medial temporal lobes. Injuries to the vessels that supply these regions can lead to a variety of syndromes depending upon the specific nature of the injury. (See 'Clinical presentation' below and "Stroke: Etiology, classification, and epidemiology" and "Posterior circulation cerebrovascular syndromes".)

The VAs most commonly originate from the subclavian arteries, although some variation can occur. The VAs originate directly from the aortic arch in 3 to 5 percent of individuals [9]. The VAs are commonly asymmetric in diameter, and it is not uncommon for one VA to be atretic, a finding that is slightly more common on the left than the right [8].

The VA traverses the neck posterior to the scalene muscle, enters the transverse foramen of the sixth cervical vertebra and exits the transverse foramen of the second cervical vertebra.

Vertebral segments — The VA is divided into four anatomic segments (V1 to V4) (figure 5).

●V1 – Origin of the vessel to the foramen of the sixth cervical (C6) transverse process.

●V2 – Intraforaminal segment from the sixth to the second cervical vertebral body (C6 to C2).

●V3 – From the second cervical (C2) foramen to the base of the skull.

●V4 – Intracerebral segment of the vertebral artery. The vertebral arteries merge to form the basilar artery and are intradural.

Site of injury — Bilateral cerebrovascular injuries are common, occurring in 18 to 25 percent of patients [2,3,10]. Carotid injuries appear to occur more frequently than vertebral injuries. In a retrospective review of 171 patients found to have blunt cerebrovascular injury, 114 patients had 157 carotid artery injuries and 79 patients had 97 vertebral artery injuries [3]. However, some studies have shown an equal number or a preponderance of vertebral injuries [11,12].

MECHANISMS OF INJURY — A variety of trauma mechanisms, including motor vehicle crashes, falls, assaults, and strangulation, can result in blunt cerebrovascular injury. Motor vehicle crashes account for more than one half of blunt cerebrovascular injuries; however, in patients 65 years and older, ground-level falls account for two-thirds of blunt cerebrovascular injury [2,3,10,13]. The absence of trauma from an individual's history does not exclude trauma as an etiology because patients often consider the inciting event to be insignificant as described below.

Blunt injury to the carotid or vertebral arteries is usually the result of a significant force that twists or stretches the vessel or impinges the vessel against the underlying bone, often for only a brief period of time. The carotid or vertebral artery may also be lacerated by bone that has fractured.

Four fundamental mechanisms of blunt carotid injuries are described [14]:

●Type I carotid injuries involve the direct application of force to the neck (eg, seat belt, strangulation, near-hanging) and account for up to approximately 10 percent of blunt carotid injuries.

●Type II carotid injuries are due to hyperextension or contralateral rotation of the head and neck (figure 6). This is probably the most common mechanism of injury. The lateral articular processes and pedicles of the upper three cervical vertebrae (C1 to C3) project more anteriorly than those of the lower four cervical (C4 to C7) vertebrae. Thus, in the upper region, the cervical internal carotid artery (ICA) can be stretched across the lateral processes with cervical hyperextension. Rotation at the atlantoaxial joint aggravates stretch injury because it can cause anterior movement of the contralateral C1 transverse process [15].

●Type III carotid injuries involve intraoral trauma that affects the internal carotid artery at the angle of the jaw (figure 6). This type of injury is seen in patients who fall with a hard object in the mouth such as in children who have fallen with a toothbrush in their mouth [16].

●Type IV injuries are due to laceration of the carotid artery resulting from basilar skull fracture, most commonly in the region of the carotid canal [17].

Regardless of the underlying mechanism of injury, the pathologic insult in most cases is an intimal tear. The exposed subendothelial collagen promotes platelet aggregation and thrombus formation, which may occlude the vessel altogether or embolize to the cerebral circulation.

The intimal tear may remain static or there may be a subintimal dissection that progresses cranially, which can cause luminal narrowing or acute vessel occlusion as the false lumen collapses against the true lumen. Less commonly, partial or complete transection of the artery occurs, resulting in pseudoaneurysm formation or free rupture. A pseudoaneurysm may increase in size to compress and occlude the vessel lumen, or rupture; if it contains thrombus, it may be a source of emboli. Rupture may result in intracranial or extracranial hemorrhage or formation of an arteriovenous fistula.

Other cerebrovascular injury mechanisms have been reported. As an example, numerous case reports document blunt injuries (mostly dissection) following mechanisms that would not typically raise a suspicion for injury, so-called "trivial trauma." These mechanisms include chiropractic manipulation, hyperflexion or hyperextension during hair washing, "head banging" to music, "bottoms-up" drinking, rapid head turning, virtually any athletic endeavor, and everyday activities such as coughing, shaving, vomiting, and nose-blowing [8].

"Spontaneous" carotid and vertebral artery dissections occur in the absence of apparent etiologic factors and account for approximately 2 to 3 percent of primary strokes [18]. Risk factors include hypertension, Marfan syndrome, fibromuscular dysplasia, syphilis, arteriopathies, and Erdheim cystic medial necrosis. Some speculate that the presence of these medical risk factors predisposes these patients to cerebrovascular injury following lesser degrees of trauma. Spontaneous carotid and vertebral artery dissections are discussed separately. (See "Cerebral and cervical artery dissection: Clinical features and diagnosis".)

TRAUMA EVALUATION — The initial resuscitation, diagnostic evaluation, and management of the patient with blunt or penetrating injury is based upon protocols from the Advanced Trauma Life Support (ATLS) program, established by the American College of Surgeons Committee on Trauma. The initial resuscitation and evaluation of the patient with blunt trauma is discussed elsewhere. (See "Initial evaluation and management of blunt abdominal trauma in adults" and "Initial evaluation and management of blunt thoracic trauma in adults" and "Initial evaluation and management of facial trauma in adults", section on 'Midface'.)

Delayed recognition of blunt cerebrovascular injury may result from the need for clinicians to manage life-threatening thoracic or abdominal injuries. In some cases, a meaningful neurologic examination is not possible due to head injuries or the need to sedate or intubate the patient.

Associated injuries — Blunt cerebrovascular injury is associated with mechanisms that can cause severe head, facial, spine, chest, and abdominal injuries [8,19]. In a review of 171 patients with blunt cerebrovascular injury, the average injury severity score was 28±1, which is consistent with other reports. Associated injuries included the brain (57 percent), spine (44 percent), chest (43 percent), and face (34 percent) [3]. The overall incidence of craniofacial fracture-associated blunt cerebrovascular injury was 0.45 percent in a systematic review, typically in association with high-energy injuries [19]. Mandibular and LeFort fractures were the most common isolated craniomaxillofacial injury associated with blunt cerebrovascular injury.

Cervical spine fracture has the strongest association with blunt cerebrovascular injury [20-24].

●In an early analysis of screening criteria, cervical spine fracture was identified as the only independent risk factor for vertebral artery injuries (odds ratio [OR] 14.50, 95% CI 5.30-39.63) [23].

●One systematic review and meta-analysis identified only cervical spine injury and thoracic injury as significantly increasing the risk for blunt cerebrovascular injury (OR 5.45, 95% CI 2.24-13.27; OR 1.98, 95% CI 1.35-2.92; respectively) [22].

●In a systematic review of studies using computed tomography angiography for screening, fractures in the C1-C3 region, two-level fractures, subluxations/dislocations, and transverse foramen fractures were associated with increased incidence of blunt cerebrovascular injury [20]:

•C1-C3: OR 2.2 (95% CI 1.1–4.2)

•Two-level: OR 2.5 (95% CI 1.4–4.6)

•Subluxation/dislocation: OR 2.9 (95% CI 1.8–4.5)

•Transverse foramen: OR 3.6 (95% CI 1.4–8.9)

Vertebral artery injuries are associated with cervical spine injuries at all levels [25]. The risk of vertebral artery injury is higher for cervical spine fractures of the upper three vertebral bodies as well as dislocations [20,26]. Other fracture patterns associated with vertebral artery injury include C1 and C2 combined fractures, transverse foramen fractures, and subluxation of adjacent vertebral levels [20,26,27]. In a review that stratified cervical spine fractures as high risk (C1 to C3) or low risk (isolated, low [C4 to C7]) for blunt cerebrovascular injury, the incidence of blunt cerebrovascular injury was 8 and 2 percent, respectively, suggesting the need to screen all cervical spine injuries [28]. A practice management guideline from the Eastern Association for the Surgery of Trauma conditionally recommends screening for low-risk cervical spine injury patterns, given that a known, albeit small, percentage of such patients will have injuries and thus be at risk of stroke [29]. (See 'Screening' below.)

CLINICAL PRESENTATION — The clinical presentation of blunt cerebrovascular injury can vary greatly depending upon the vessel affected, site of injury, injury grade, and any preexisting cerebrovascular disease. (See 'Cerebrovascular anatomy' above and "Blunt cerebrovascular injury: Treatment and outcomes", section on 'Injury grading'.)

The majority of patients with blunt cerebrovascular injury have no obvious neurologic manifestations at presentation [3]. A latent period between the time of injury and the appearance of clinical manifestations is typically seen. Unless the vessel immediately occludes, time is required for thrombus formation that might limit flow or lead to distal embolization. In various studies, 25 to 50 percent of patients first developed signs or symptoms of blunt carotid injury more than 12 hours after the traumatic event [2,4,30-32].

The following observations have been made in studies in which screening for cerebrovascular injury was performed in patients with blunt trauma:

●In 17 patients diagnosed with blunt cerebrovascular injury identified from 3480 blunt trauma admissions, 10 deaths were attributable to the injury [4]. The median time until diagnosis was 12.5 hours for the entire group and 19.5 hours for nonsurvivors. Four of five patients whose diagnosis was delayed for more than 48 hours did not survive.

●In an analysis of 45 blunt cerebrovascular injury-related strokes occurring in patients who could not receive antithrombotic therapy, 11 out of 45 occurred within two hours of injury; the average time to stroke in the remaining 34 of 45 patients was 75 hours [31].

●In another retrospective review of 76 patients with blunt cerebrovascular injury, 42 percent of symptomatic patients manifested symptoms more than 18 hours following injury, with two patients becoming symptomatic at seven days [32].

●In a multicenter review, among patients with blunt cerebrovascular injury-related stroke, 63 percent had no signs at initial presentation of the injury; stroke was identified at a median 48 hours after admission [33]. Among those without reliable neurologic exam, the median time to imaging diagnosis of stroke was 42 hours; among those who developed signs or symptoms, the median time was 53 hours.

●The average time to stroke can range from two hours to one week, but most occur between 12 and 75 hours [2,4,30-32].

When neurologic manifestations of cerebral ischemia occur, the symptoms and findings on neurologic examination depend upon the specific artery involved, the presence (or absence) of adequate collateral circulation (figure 3), and the presence (or absence) of underlying cerebrovascular disease. Approximately 10 to 15 percent of symptomatic patients have lateralizing symptoms, and at least 50 percent develop a completed stroke. No single neurologic finding allows a precise diagnosis, but the constellation of findings may identify the involved artery [8]. The manifestations of cerebral ischemia are reviewed elsewhere. (See "Overview of the evaluation of stroke" and "Definition, etiology, and clinical manifestations of transient ischemic attack" and "Stroke: Etiology, classification, and epidemiology" and "Posterior circulation cerebrovascular syndromes".)

Certain clinical signs are so suggestive of blunt cerebrovascular injury that they should prompt emergent diagnostic evaluation, usually with computed tomography (CT) (see 'Imaging evaluation' below). These include acute arterial hemorrhage from the neck, mouth, nose, or ear; expanding cervical hematoma; cervical bruit in a patient younger than 50 years of age; and focal or lateralizing neurologic deficits (eg, hemiparesis, transient ischemic attack, Horner syndrome, oculosympathetic paresis, vertebrobasilar insufficiency) (algorithm 1). Patients with acute hemorrhage are managed to limit the amount of blood loss and prevent shock. (See "Treatment of severe hypovolemia or hypovolemic shock in adults" and "Blunt cerebrovascular injury: Treatment and outcomes", section on 'Injury grading'.)

Symptoms such as neck, ear, face, or periorbital pain occur in up to 60 percent of patients with carotid or vertebral artery dissection in the neck [34]. However, pain may be difficult to elicit in the multiply injured patient or may be attributed to other injuries. Thus, any patient with neck pain or headache following blunt trauma should be suspected of having cerebrovascular injury.

Neurologic signs associated with local arterial injury, which occurs in approximately 5 percent of symptomatic patients, suggest that blunt cerebrovascular injury has occurred even though patients have no manifestations of cerebral ischemia. As an example, blunt carotid injury can disrupt the periarterial sympathetic plexus, leading to Horner syndrome (ptosis, myosis, anhidrosis) or oculosympathetic paresis, which is a partial Horner syndrome that includes ptosis and miosis. Thus, although pupillary asymmetry can have a number of etiologies in the trauma patient, blunt cerebrovascular injury should be suspected ipsilateral to the side of the smaller pupil if the larger of the pupils is reactive and the smaller pupil is not [35]. (See "Horner syndrome".)

SCREENING — We agree with trauma guidelines that recommend screening patients with high-risk cervical injuries [29,36]. Selected patients with low-risk injuries may also benefit from screening [29,37,38]. As described in the preceding section, a majority of patients with blunt cerebrovascular injury do not have neurologic symptoms at presentation but may develop signs or symptoms hours or days later. This delay provides a window of opportunity during which patients at risk for adverse outcomes can be identified and stroke prevention therapy initiated [1,39-41]. (See 'Clinical presentation' above.)

Failure to identify and treat blunt cerebrovascular injury when symptoms occur results in significant morbidity and mortality, but it is unclear whether aggressive screening leads to improved outcomes [42,43]. (See "Blunt cerebrovascular injury: Treatment and outcomes", section on 'Stroke and mortality'.)

When screening was first proposed for blunt cerebrovascular injury, it was controversial primarily because the only reliable test available was catheter-based arteriography. Moreover, treatment with systemic anticoagulation in acutely injured patients based upon positive screening findings was thought to be too risky. Screening has clearly increased the number of blunt cerebrovascular injuries being diagnosed [3,5,11,40,44-46]. In the earliest study of screening, 171 patients with indications to exclude thoracic aortic injury using digital subtraction arteriography also underwent carotid arteriography. Although these patients had no symptoms or signs to suggest blunt cerebrovascular injury, 3.5 percent of the studies were positive [47]. Subsequently, using defined clinical criteria, blunt cerebrovascular injury was identified in 27 percent of the asymptomatic patients screened [23]. After instituting a policy of screening, the incidence of blunt carotid injury increased from 0.1 to 1.6 percent [1,3]. In separate review of 8292 patients screened for blunt cerebrovascular injury, 47 (0.5 percent) had blunt vertebral artery injury [48]. These findings have been replicated at a number of trauma centers [5,11,40,42,44,45].

Some investigators have suggested that blunt cerebrovascular injury cannot be reliably predicted based upon the injury mechanism or injury patterns, and thus advocate for universal screening. However, others have critically evaluated specific injury mechanisms, patterns of injury, and clinical symptoms and signs of blunt cerebrovascular injury to define criteria by which to perform diagnostic testing.

●A multivariate analysis identified independent risk factors for blunt carotid injury that included Glasgow coma scale (GCS) <6 (table 1), petrous bone fracture, diffuse axonal brain injury, and LeFort II or LeFort III fracture (figure 7) [23]. If one of these high-risk factors was present, the patient had a 41 percent chance of having an injury, whereas if all four factors were present, the likelihood of injury increased to 93 percent. On the other hand, 20 percent of patients who had blunt cerebrovascular injury did not have any of these risk factors. The only independent predictor of blunt vertebral artery injury was cervical spine fracture. (See 'Indications for imaging' below.)

●Another retrospective review used the above criteria to screen patients for blunt cerebrovascular injury and included additional criteria to capture specific cervical spine fracture patterns [41]. Of those who met the screening criteria, 244 patients had a blunt cerebrovascular injury for an incidence of 1.5 percent. Carotid and vertebral injuries were seen nearly equally (141 carotid, 124 vertebral). More than one injury was seen in 37 percent of the patients.

The Denver criteria for screening, which are based on the outcomes of these studies, are summarized in the algorithm (algorithm 1). A study supporting the "expanded Denver Criteria" reported that 11 percent of patients with blunt cerebrovascular injury did not fulfill standard screening criteria [38].

While many trauma centers have embraced screening, the optimal screening criteria continue to be debated [22,25,26,40,49-51]. More restrictive criteria will yield a higher proportion of positive results and lower cost, but an increase in the likelihood of missing an injury. With the use of noninvasive studies for screening, particularly CT angiography, the criteria used for screening have been broadened [52]. Still, investigators are reporting that 30 to 37 percent of patients with blunt cerebrovascular injuries have none of the reported clinical or radiographic risk factors [46,53]. It has been suggested that noninvasive screening for blunt cerebrovascular injury be performed in any patient who has sustained a mechanism of injury sufficient to warrant either a CT of the cervical spine or a CT angiography of the chest [53]. (See 'Imaging evaluation' below.)

Two centers have suggested protocols for universal screening. While implementation is likely to yield more injuries, the additional costs and potential harms need to be considered along with the mechanism of injury.

●In one study, 4659 blunt trauma patients underwent screening CT angiography, with 126 (2.7 percent) found to have blunt cerebrovascular injury [54]. Of the 51 percent who met screening criteria, 83 percent had injuries. All 14 grade III to V injuries in patients who did not meet screening criteria were involved in motor vehicle crashes (MVCs).

●In another study, 6287 blunt trauma patients underwent screening CT angiography, with 480 (7.6 percent) found to have blunt cerebrovascular injury [55]. Twenty-five percent did not meet screening criteria. Fifty percent of the injuries were grade I injuries; 64 percent of patients were injured in MVCs.

Cost effectiveness — Screening for blunt cerebrovascular injury in patients with risk factors appears to be cost effective. In one analysis, CT angiography was the most cost-effective screening strategy, preventing the most strokes at a reasonable cost [56].

A Markov decision analysis model was used to compare various screening strategies, from no screening to universal screening [57]. The authors concluded that universal screening was the optimal strategy when the incidence of blunt cerebrovascular injury was greater than 6 percent. However, an incidence rate this high has only been reported in one study [55], so further research is required.

INDICATIONS FOR IMAGING — Neurovascular imaging is indicated for patients with any clinical symptom or sign suggestive of blunt cerebrovascular injury, patients with unexplained neurologic symptoms, and in asymptomatic patients with any of the risk factors listed below (algorithm 1) [22,50,58].

The following clinical signs are suggestive of blunt cerebrovascular injury in the trauma patient and should prompt emergent evaluation and interventions directed at hemorrhage control or stroke management. (See 'Clinical presentation' above.)

●Potential arterial hemorrhage from the neck, mouth, or ear.

●Expanding cervical hematoma.

●Cervical bruit in a patient younger than 50 years of age.

●Focal or lateralizing neurologic deficit (eg, hemiparesis, transient ischemic attack, Horner syndrome, oculosympathetic paresis, vertebrobasilar insufficiency).

●Neurologic deficit inconsistent with head computed tomography (CT).

●Stroke on CT or magnetic resonance imaging.

The following risk factors for blunt cerebrovascular injury have been endorsed by both the Eastern and Western Trauma Associations of the United States to screen asymptomatic patients (algorithm 1) [29,36,59,60]:

●Injury mechanism compatible with severe cervical hyperextension/rotation or hyperflexion.

●Severe facial trauma including bilateral facial fractures in any facial third, complex midface (figure 7), and subcondylar fractures.

●Basilar skull fracture involving the carotid canal.

●Closed head injury consistent with diffuse axonal injury with Glasgow Coma Score <6.

●Cervical vertebral body or transverse foramen fracture, subluxation, or ligamentous injury at any level, or any fracture at the level of C1 to C3.

●Near hanging resulting in cerebral anoxia.

●Clothesline-type injury or seat belt abrasion associated with significant cervical pain, swelling, or altered mental status.

●The presence of major thoracic trauma has been recommended as a screening criterion by a number of groups [22,50]. This includes traumatic brain injury in association with thoracic injuries, thoracic vascular injuries, and blunt cardiac rupture.

●Upper rib fractures (ribs 1 to 3; often associated with major thoracic injury).

IMAGING EVALUATION — Patients with symptoms, signs, or risk factors for injury are presumed to have blunt cerebrovascular injury until proven otherwise. (See 'Indications for imaging' above.)

Imaging modalities

Digital subtraction arteriography — Four-vessel biplanar digital subtraction arteriography (DSA) was the original study used for screening and remains the criterion standard for the diagnosis of blunt cerebrovascular injury. Each of the four vessels (bilateral carotid, bilateral vertebral) should be imaged from their origins in the chest, and cerebral arteriography is included to identify the presence of occluded terminal vessels and evaluate collateral circulation to the brain (image 1) Arteriography is invasive and associated with access site complications (eg, hematoma, pseudoaneurysm), contrast nephropathy, contrast allergy, and stroke [3,61,62]. In addition, DSA is expensive and may not be uniformly available. Consequently, noninvasive tests have largely supplanted DSA for screening.

CT angiography — Computed tomographic (CT) angiography has become the screening test of choice for blunt cerebrovascular injuries (algorithm 1). This is primarily because patients who are candidates for screening often have indications for scanning other regions of the body such as the head, face, cervical spine, chest, and abdomen. Another benefit of CT is a reduction in the volume of contrast needed to perform these studies, which can be up to one-half that needed for arch aortography and four-vessel cerebral arteriography. (See 'Trauma evaluation' above and 'Indications for imaging' above.)

The major concern regarding CT arteriography and other noninvasive imaging studies is accuracy. Early-generation (<16 multidetector-row) CT scanners had inadequate sensitivity and specificity for blunt cerebrovascular injury, but that has increased with improvements in CT technology [63].

●One study screened 146 patients for blunt cerebrovascular injury with both arteriography and CT angiography [11]. Injuries were identified in 46 patients (20 carotid, 26 vertebral). The sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of CT angiography were 98, 100, 100, 99, and 99 percent, respectively. The single false positive result occurred in a patient with a grade I vertebral artery injury.

●Another study screened 372 blunt trauma patients with 16-slice CT angiography identifying 271 normal studies [64]. Arteriography was performed on a subset (82 patients) with normal CT angiography and identified seven missed blunt cerebrovascular injuries. However, on re-review of the CT images, the injuries were evident in six of the seven patients; the seventh patient's abnormality was most likely not traumatic in origin. The inaccuracy of CT angiography appeared to be related, in part, to the radiologists' experience, as all of the missed blunt cerebrovascular injury occurred in the first half of the study period. The accuracy of CT may also be related to the grade of injury.

●A systematic review pooled the results of eight studies for CT angiography (1-, 4-, 8-, 16-, 32-, and 64-slice scanners) using DSA as the standard [65]. The pooled sensitivity and specificity for blunt cerebrovascular injury for CT angiography was 66 and 97 percent, respectively. Sensitivities varied with the number of available CT slices, the training of interpreting radiologists, and the diagnostic threshold for judging a study as positive. The sensitivity remained ≤80 percent among studies that used scanners with 16 or more slices and where the CT angiogram was read by neuroradiologists.

●A retrospective review reported a positive predictive value of CT angiography of only 30 percent for grade I injuries, compared with 76 percent for grade II and 97 percent for grade IV injuries [66]. The overall rate of false-positive CT angiography for blunt cerebrovascular injury was 48 percent for patients screened.

Although CT angiography may lack sensitivity to completely rule out blunt cerebrovascular injury, it remains the best screening study to rule in blunt cerebrovascular injury among trauma patients with a high pretest probability of injury (ie, appropriate risk factors) [67]. Using a 16-slice scanner, the detected incidence of blunt cerebrovascular injury is consistent with that seen using DSA [44,45]. Whole-body multidetector-row CT has now been adopted in some centers. It offers more rapid imaging with a single dose of intravenous contrast. A preliminary report suggests an accuracy equivalent to 16-slice CT angiography, but further evaluation is needed [68].

MR angiography — Potential advantages of magnetic resonance (MR) include the avoidance of iodinated contrast agents, lack of bony artifact, and earlier detection of cerebral infarction (image 2A-C). (See "Neuroimaging of acute stroke".)

A number of studies have described the use of MR angiography to identify blunt cerebrovascular injury [69-74]. However, when compared with catheter-based arteriography, studies have reported poor specificity at 67 percent, with sensitivities ranging from 50 to 75 percent [5,63,75]. A meta-analysis reported overall sensitivity of 55 percent [76]. Given issues such as the time required for the study and incompatibility with implanted devices, MR angiography is not used as an independent primary modality for screening blunt cerebrovascular injury [77].

Magnetic resonance imaging (MRI) can help characterize high-risk blunt cerebrovascular injury that can lead to stroke. In a review of 145 affected vessels from 118 patients, the presence of irregular surface and intraluminal thrombus on imaging was independently associated with acute ischemic stroke (odds ratio [OR] 4.29 95% CI, 1.61–11.46; OR 7.48 95% CI, 1.64–34.07; respectively) [78].

MRI can also be used as a problem-solving tool in differentiating anatomic variants from clinically significant abnormalities by demonstrating acute perivascular inflammatory edema or differentiating acute from chronic intramural hematoma. High-resolution vessel wall imaging and MR angiography sequences, such as double inversion recovery black-blood imaging, are reliable in demonstrating intimal flap and intramural hematoma in small subsets of patients [79-81].

Duplex ultrasonography — Duplex ultrasonography was the first noninvasive study considered for the evaluation of blunt cerebrovascular injury [82], and although it is the test of choice for evaluating other pathologies (eg, atherosclerosis) of the extracranial carotid arteries, the data do not support ultrasound as an appropriate screening modality for blunt cerebrovascular injury [59].

Duplex sonography is not a highly sensitive test to screen for blunt cerebrovascular injury primarily because imaging is limited to the cervical portions of the vessels; the majority of these injuries, particularly hyperextension/flexion injuries, occur at the base of the skull [2,83-85]. Although duplex ultrasound can provide indirect evidence of distal injuries (eg, obstructive waveforms proximal to a flow limiting lesion), indirect findings are not reliable when the degree of stenosis is less significant (ie, <60 percent). Thus, grade I, grade III, and many grade II injuries could potentially be missed. (See "Blunt cerebrovascular injury: Treatment and outcomes", section on 'Injury grading'.)

In addition, patients with skin abrasion and soft tissue swelling may not tolerate the examination due to pain. In addition, removal of the cervical collar is required to perform the examination. Given that these patients are at risk for cervical spine fracture, collar removal is often contraindicated.

Duplex ultrasonography may have a limited role in evaluating select cases that involve direct blows to the anterior neck, follow-up of proximal lesions, and possibly in children, in whom other imaging techniques might require conscious sedation. When ultrasound is used, the ultrasonographer must be skilled, and the interpreting clinician must recognize the limitations of the study in the trauma population.

Choice of imaging — Catheter-based DSA was historically used to diagnose blunt cerebrovascular injury; however, the use of noninvasive imaging predominates for the reasons discussed in the sections above. CT angiography (≥16 slice) is the most reliable noninvasive screening modality for blunt cerebrovascular injury. A relatively high false positive rate is reported in some series, and thus, follow-up DSA may be warranted to definitively exclude an injury, or provide a diagnosis when a high clinical suspicion remains but the CT study is normal. (See 'CT angiography' above.)

As discussed above, the sensitivity for CT angiography for blunt cerebrovascular injury depends upon the number of scanner slices and the experience of the interpreting radiologist. Thus, the evaluation of blunt cerebrovascular injury at a given institution will depend upon the available resources; DSA should be used for high-risk patients if an appropriately sensitive CT scanner is not available. If DSA or high-quality CT scanning are not available, the patient should be transferred to a trauma center for further evaluation .

In our opinion, the role for DSA is to make a definitive diagnosis in a symptomatic patient with equivocal or normal-appearing CT angiogram, or to plan interventions (algorithm 1). Some clinicians have suggested a more liberal use of DSA to confirm positive findings on CT to reduce the likelihood that a patient would be unnecessarily anticoagulated [86]. (See 'Digital subtraction arteriography' above.)

Duplex ultrasonography may play a role in the long-term follow-up of cervical carotid injuries; however, duplex ultrasound and magnetic resonance imaging are not sufficiently sensitive to be used for screening. (See 'Duplex ultrasonography' above and 'MR angiography' above.)

Injury grading — Grading the injury based upon its appearance on imaging was created to standardize clinical communication and guide therapy [32]. The treatment of blunt cerebrovascular also depends, in part, on the grade of injury. Injury grading is discussed in detail elsewhere. (See "Blunt cerebrovascular injury: Treatment and outcomes", section on 'Injury grading'.)

Follow-up imaging — A significant number of injuries evolve in a manner that could alter therapy, and follow-up imaging is warranted in patients with blunt cerebrovascular injury (algorithm 1). The choice and timing of follow-up imaging is discussed elsewhere. (See "Blunt cerebrovascular injury: Treatment and outcomes", section on 'Follow-up for healing or progression'.)

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Blunt cerebrovascular injury".)

SUMMARY AND RECOMMENDATIONS

●Blunt cerebrovascular injury – Blunt carotid and vertebral artery injuries are collectively termed blunt cerebrovascular injury and are rare but potentially disastrous injuries. The overall incidence of these injuries is approximately 1-3 percent of all blunt trauma admissions. (See 'Introduction' above.)

●Cerebrovascular anatomy – The circulation to the brain originates from the carotid and vertebral arteries comprising the anterior and posterior circulations, respectively. The circle of Willis, which provides communication between the anterior and posterior circulation, is completely intact and symmetric in only approximately 20 percent of individuals. Because of the inconsistency, the neurologic presentation of patients with blunt cerebrovascular injury is highly variable. (See 'Cerebrovascular anatomy' above.)

●Clinical features and diagnosis – Patients sustaining blunt trauma who have the following clinical symptoms or signs should undergo urgent/emergency imaging to evaluate for cerebrovascular injury (algorithm 1):

•Potential arterial hemorrhage from the neck, mouth, nose, or ear

•Expanding cervical hematoma

•Cervical bruit in patients <50 years old

•Focal or lateralizing neurologic deficits

•Neurologic deficit inconsistent with head computed tomography (CT) or magnetic resonance imaging (MRI)

•Stroke on CT or MRI

●Screening asymptomatic patients – We screen for blunt cerebrovascular injury in asymptomatic patients who have the risk factors for blunt cerebrovascular injury listed below (algorithm 1). We prefer to use multislice CT (≥16 slice) for screening. If an appropriately sensitive CT scanner is not available, catheter-based digital subtraction arteriography (DSA) should be performed. For institutions without this imaging capability, transfer to a trauma facility is indicated. (See 'Indications for imaging' above and 'Choice of imaging' above.)

•High-energy transfer mechanism

•Cervical hyperextension/hyperflexion injury

•Severe facial trauma including bilateral facial fractures in any facial third, complex midface (figure 7), and subcondylar mandibular fractures

•Complex skull fracture/basilar skull fracture/occipital condyle fracture

•Severe traumatic brain injury (TBI) with Glasgow Coma Score <6

•Cervical spine fracture, subluxation, or ligamentous injury at any level

•Near-hanging resulting in cerebral anoxia/strangulation, especially with loss of consciousness

•Clothesline-type injury or seat belt abrasion associated with significant cervical pain, swelling, or altered mental status

•TBI with thoracic injuries

•Scalp degloving

•Thoracic vascular injuries

•Blunt cardiac rupture

•Upper rib fractures (ribs 1 to 3)