Blunt cerebrovascular injury (BCVI) is characterized by nonpenetrating trauma to the vertebral and/or internal carotid arteries resulting in an intraluminal or intramural hematoma that leads to stenosis or occlusion.4,24 The incidence of BCVI varies widely but is generally considered to involve between 1% and 9% of pediatric patients seen in the emergency department for blunt trauma–related injuries.12,20,24,26,27 If not diagnosed early, BCVI could lead to significant neurological deficits and possibly death.3,20,47
While BCVI has been well described in the adult population, there is a paucity of research regarding this entity in children. Currently, there is no standardized treatment for pediatric BCVI, and the risk factors for pediatric BCVI have not been clearly described.42 Previous studies have suggested that the risk factors in the pediatric population may mimic those in adults.35 However, recent research shows that the patterns of intracranial injury in pediatric trauma patients as observed on head CT scans differ significantly from those seen in the adult trauma population.43 Additionally, treatment of BCVI varies in the adult and pediatric populations. Currently, the most common treatment for adult BCVI is anticoagulation therapy. Administering antithrombotic or anticoagulation therapy soon after injury may help patients who require it regain neurological function and minimize mortality.17,25 However, anticoagulation also carries a significant risk of hemorrhage, especially in the setting of coincident intracranial hemorrhage.1 Endovascular stents have also been placed in select adult patients; however, stents carry the risk of hemorrhagic complications.15 The treatment for pediatric BCVI is even less standardized than that for adult BCVI, and treatment protocols vary widely by trauma center.6 Some studies have suggested that operative intervention is not associated with increased survival, and medical management might be the best way to treat low-grade pediatric BCVI injuries.36,44,47 For this reason, it is imperative that research be conducted to develop a reliable method for screening pediatric BCVI patients and to develop a standardized method that institutions can use to evaluate BCVI risk. Only by developing an accurate screening tool can a consensus be drawn regarding the best approach for treating pediatric BCVI.
One of the difficulties with diagnosing BCVI is that the symptoms might be masked by other trauma to the head and neck. Furthermore, many patients do not initially present with focal neurological findings.3,47 For these reasons, and the lack of sufficient research, the true incidence of BCVI in pediatric patients is unclear. As a result, some have proposed a more inclusive screening protocol to identify BCVI with imaging, such as CT angiography (CTA), MR angiography (MRA), and angiography.5,9,33 This is similar to the liberal screening of adult patients who present to the emergency department with possible BCVI.5,33 While some studies have demonstrated improved BCVI detection using angiographic imaging in the adult population,45 there is direct evidence that the organ dose that corresponds to a common CT study results in an increased risk of cancer in children.23 Children are considerably more sensitive to the carcinogenic effects of ionizing radiation and have a longer life expectancy during which the cancer risk can accumulate and manifest; as a result, no amount of radiation is safe in children.11,34 Studies have shown that 50 mGy of radiation exposure may triple the risk of leukemia, radiation doses of 60 mGy might triple the risk of brain cancer,40 and dose-reduction strategies could dramatically reduce the number of radiation-induced cancers.39 This risk is further compounded in pediatric patients at risk for BCVI because diagnosing BCVI requires angiographic imaging such as CTA, which contains an estimated radiation dose of 12–16 mSv.2,29,31 This is equivalent to 120% of the dose of an abdominal CT study or 600 chest radiographs.29 These risks indicate that a good screening tool is needed to accurately determine which pediatric patients are at high risk of BCVI in order to minimize unnecessary radiation exposure and future cancer incidence.
Currently there are 2 primary adult screening criteria for determining whether patients who present to the emergency department with BCVI need angiographic imaging: the Denver screening criteria and the modified Memphis criteria.6 However, wide variation exists in the literature regarding the effectiveness of each of these criteria. A large systematic review and meta-analysis of more than 120,000 patients determined that BCVI was not identified in 20%–22% of adult patients using the Denver screening criteria.27 Similarly, the modified Memphis criteria was shown to miss 37.5% of adult trauma injuries in a large retrospective study.10 Due to the lack of literature discussing BCVI in pediatric patients, the Eastern Association for the Surgery of Trauma (EAST) looked at BCVI risk factors in a pediatric trauma cohort and determined that the risk factors for pediatric BCVI are similar to those in the adult population. Thus, EAST recommended a screening score for the pediatric trauma population that is very similar to those used for adults.12,35 Subsequently, several studies attempted to validate the use of adult screening criteria in pediatric BCVI, including the Denver group, which found that only 30% of symptomatic children with BCVI met their previously reported adult screening criteria.32 Recently, however, Ravindra et al. at the University of Utah School of Medicine sought to retrospectively design a score that is specifically tailored toward identifying high-risk BCVI patients in a large pediatric trauma cohort. They reported demographic, clinical, and imaging characteristics in a cohort of pediatric patients diagnosed with BCVI to identify variables that could be used to determine which pediatric trauma patients are at a higher risk for developing BCVI.41,42 Additionally, these authors proposed a screening tool based on clinical evaluations and CT imaging to identify which patients who present to the emergency department with blunt trauma should undergo angiography in order to make a definitive diagnosis of BCVI. Although this score was validated in pooled data from other centers (Monroe Carell Children’s Hospital, Nashville, Tennessee; St. Louis Children’s Hospital, St. Louis, Missouri; Texas Children’s Hospital, Houston, Texas), the sensitivity remained low at 59%, thereby indicating the questionable utility of this score as an initial screening tool.42
This study is unique in that we incorporated the mechanism of injury (MOI) into our screening score, which is a criterion that was not previously examined by any of the aforementioned screening protocols. We used the Utah score as a comparison metric for our own screening tool, the McGovern score, because the former was the only screening tool specifically designed for the pediatric population. We evaluated stroke after BCVI in this population and assessed anatomical considerations in the vessel injury as well. We hope that by evaluating independent BCVI risk factors in conjunction with the Utah score at our Level 1 pediatric trauma center, we can develop a better screening tool to identify pediatric patients at high risk for BCVI and minimize unnecessary radiation exposure. Additionally, we hope to assess the incidence of BCVI in pediatric trauma patients as well as retrospectively evaluate the risk factors and treatment regimens for patients diagnosed with BCVI at our institution.
Methods
Patient Population
This hospital-based cohort study was conducted in Houston, Texas, at Memorial Hermann Hospital, which is a Level 1 pediatric trauma center, of pediatric patients who presented to our institution with BCVI between 2005 and 2015. Approval was obtained from the UT Health Committee for the Protection of Human Subjects. A trauma registry was compiled in which all pediatric patients (< 16 years) who were seen in the emergency department and admitted to the intensive care unit for trauma were prospectively recruited over the 10-year period of study. In total, 12,614 patients were recruited as part of this trauma registry. Patients (n = 44) who sustained penetrating injuries (i.e., gunshot wound or stab injury) to the head or neck were excluded from this study. Next, the data set was queried to identify trauma patients who underwent angiography (CTA, MRA, or angiography of the head and/or neck). In total, 460 patients were identified. These 460 patients were then individually graded for BCVI as determined by the Biffl scale, the criteria of which are used to classify the extent of BCVI injury based on angiographic imaging.7 Twenty-one patients were found to have BCVI.
Data Collection
Among the patients who underwent angiography (n = 460), the imaging modality, patient age, arrival date, length of follow-up, and MOI were recorded. The MOI was categorized as a motor vehicle accident, automobile-pedestrian accident, bicycle accident, fall, or other. The clinical information recorded for each patient who underwent angiography included Glasgow Coma Scale (GCS) on arrival and the presence or absence of a focal neurological deficit. The obtained radiological variables included the presence or absence of a carotid canal fracture, petrous temporal bone fracture, and cerebral infarction as determined on CT scanning. Patients who underwent angiography were then individually evaluated. Records of those found to have BCVI (n = 21) were queried for mode of treatment, type of intracranial injury, artery damaged, and BCVI injury grade, in addition to the aforementioned clinical and radiological variables mentioned. Additionally, Le Fort fractures, cervical spine fractures, cervical spine subluxation, and neck soft-tissue injuries (e.g., seatbelt sign, hypoxia as a result of squeezed arteries, hematoma) were noted in the BCVI cohort (n = 21). The mode of treatment was characterized as observation, antiplatelet therapy, anticoagulation therapy, endovascular stenting, or open surgery. The type of intracranial injury was characterized as epidural, subdural, subarachnoid, or contusion. The artery damaged was characterized as intracranial carotid, extracranial carotid, intracranial vertebral, or extracranial vertebral. Lastly, the BCVI injury grade was classified according to the Biffl grading scale.7 In this method of classification, a grade I injury involved intimal irregularity with < 25% narrowing, a grade II injury involved dissection of a vessel or the presence of an intramural hematoma with > 25% narrowing, a grade III injury involved the presence of pseudoaneurysm, a grade IV injury was characterized by an occlusion, and a grade V injury involved the transection of the vessel with extravasation.7
Statistical Analysis
The data collected are summarized using the means and standard deviations for continuous variables and counts and frequencies for categorical variables. The recorded categorical variables were compared using Fisher’s exact test or chi-square test.
To assess the Utah score’s ability to predict BCVI incidence in this study’s cohort, the Utah score was calculated for all patients who underwent angiography (n = 460). The variables involved in the Utah score included GCS score ≤ 8, focal neurological deficit, carotid canal fracture as seen on CT, petrous temporal bone fracture as seen on CT, and cerebral infarction as seen on CT (Table 1). Each of the variables was weighted as 1, 2, 2, 3, and 3 points, respectively (Table 1). A cumulative score of ≥ 3 classified patients as high risk according to the Utah score34 and merited the need for angiography to confirm or rule out a diagnosis of BCVI. Once these 5 variables were evaluated in the imaging cohort, patients were categorized into a 2 × 2 table (positive/negative Utah test vs the presence/absence of BCVI) to assess the validity of the Utah score. The sensitivity, specificity, positive predictive value, and negative predictive value were then calculated (Table 2).
Utah and McGovern score criteria
| Variable | No. of Points |
|---|---|
| Utah score | |
| GCS score ≤8 | 1 |
| Focal neurological deficit | 2 |
| Carotid canal fracture | 2 |
| Petrous temporal bone fracture | 3 |
| Cerebral infarction on CT | 3 |
| McGovern score | |
| GCS score ≤8 | 1 |
| Focal neurological deficit | 2 |
| Carotid canal fracture | 2 |
| MOI | 2 |
| Petrous temporal bone fracture | 3 |
| Cerebral infarction on CT | 3 |
A score ≥ 3 points on both scales signifies high risk for BCVI and indicates that the patient should undergo angiography.
Calculation of the validity of the Utah score and McGovern score
| Scoring System | No. of Patients | PPV/NPV | Sensitivity/Specificity | |
|---|---|---|---|---|
| BCVI | No BCVI | |||
| Utah score (≥3 points) | ||||
| Positive Utah test* | 11 | 38 | 22.4%/97.6% | 52.4%/91.3% |
| Negative Utah test | 10 | 401 | ||
| Utah score (≥1 point) | ||||
| Positive Utah test† | 17 | 164 | 9.4%/98.6% | 81.0%/62.6% |
| Negative Utah test | 4 | 275 | ||
| McGovern score (≥3 points) | ||||
| Positive McGovern test‡ | 17 | 126 | 11.9%/98.7% | 81.0%/71.3% |
| Negative McGovern test | 4 | 313 | ||
NPV = negative predictive value; PPV = positive predictive value.
Refers to patients with a Utah score ≥ 3 points.
Refers to patients with a Utah score ≥ 1 point.
Refers to patients who had a score of ≥ 3 points when the MOI was taken into consideration.
To develop a more sensitive screening tool to identify BCVI, logistic regression models were generated to evaluate the sensitivity and specificity of the Utah score at varying thresholds as well as to determine which discrimination threshold was most appropriate (Fig. 1). The most predictive threshold was determined by calculating the area under the receiver operating characteristic curve (AUC) for the different Utah scores. Based on these results, a Utah score ≥ 1 was redefined as high risk, and patients with a Utah score of < 1 were redefined as low risk. The sensitivity, specificity, positive predictive value, and negative predictive value were recalculated (Table 2).
Summaries of the sensitivity and specificity of different Utah score thresholds (range 0–10) for all patients who underwent angiography. The blue curve is for the model that used the McGovern score, whereas the red curve is for the model that used the Utah score (≥ 1) alone without MOI. The AUC is shown in parentheses. ROC = receiver operating characteristic. Figure is available in color online only.
To improve the specificity (i.e., reduce the number of false positives) of the redefined score, the MOI was incorporated into the screening criteria. The new McGovern score (Table 1) assigned 2 points to patients involved in a motor vehicle accident or automobile-pedestrian accident and redefined “high-risk” patients as those patients who had a McGovern score ≥ 3. A dichotomized score was kept as suggested in the study conducted by Ravindra et al.41 to facilitate easy clinical decision making regarding whether angiography was necessary. A logistic regression model was made for the McGovern risk categorization and its association with BCVI. Using this model, the AUC for the dichotomized McGovern score was calculated (Fig. 1). The sensitivity, specificity, positive predictive value, and negative predictive value were again recalculated (Table 2). Additionally, the 21 patients diagnosed with BCVI were also evaluated using the Denver screening criteria, modified Memphis criteria, and EAST screening criteria to determine the sensitivity of each of these screening tools in our cohort. All statistical testing was performed with the R software package (v3.1.2 with the MASS library).
Results
Patient Population Characteristics
A total of 12,614 pediatric patients were prospectively enrolled in the trauma registry. Of those patients, 44 patients were excluded due to penetrating injuries (gunshot would or stab wound), and 460 patients underwent angiography (CTA, MRA, or digital subtraction angiography). Angiography was performed in all patients within 24 hours of admission, often in the emergency department. Of those patients who underwent angiography (mean age 8.1 years), 21 were found to have sustained BCVI (mean age 10.4 years), leaving 439 who were not found to have sustained BCVI. Between the 2 cohorts, patients diagnosed with BCVI had a higher incidence of motor vehicle and automobile-pedestrian accidents. In the cohort of patients who underwent angiography and were not diagnosed with BCVI (n = 439), 209 patients (47.6%) were involved in a motor vehicle accident, 37 patients (8.4%) were in an automobile-pedestrian accident, 64 patients (14.6%) sustained injuries due to a fall, 12 patients (2.7%) were involved in a bicycle accident, and 117 patients (26.7%) were injured by some other cause of trauma. In the cohort of patients diagnosed with BCVI (n = 21), 11 patients (52.4%) were involved in a motor vehicle accident, 5 patients (23.8%) were involved in an automobile-pedestrian incident, 2 patients (9.5%) sustained injuries due to a fall, 1 patient (4.8%) was injured due to a bicycle accident, and 2 patients (9.5%) were injured due to some other cause of trauma (Table 3).
Clinical and imaging characteristics in 21 patients diagnosed with BCVI
| Variable | Value |
|---|---|
| Patient characteristics | |
| Cohort size | 21 (100) |
| Mean age (SD), yrs | 10.4 (5.0) |
| Male | 12 (57.1) |
| Mean follow-up (SD), mos | 7.5 (10.9) |
| Imaging technique | |
| CTA | 21 (100) |
| MRA | 6 (28.6) |
| Digital subtraction angiography | 2 (9.5) |
| MOI | |
| Motor vehicle accident | 11 (52.4) |
| Automobile-pedestrian accident | 5 (23.8) |
| Fall | 2 (9.5) |
| Bicycle accident | 1 (4.8) |
| Other | 2 (9.5) |
| Intracranial injury | |
| Epidural hematoma | 1 (4.8) |
| Subdural hematoma | 7 (33.3) |
| Subarachnoid hemorrhage | 11 (52.4) |
| Contusion | 4 (19.0) |
| Artery injured | |
| Intracranial carotid | 7 (33.3) |
| Extracranial carotid | 9 (42.9) |
| Intracranial vertebral | 0 (0) |
| Extracranial vertebral | 5 (23.8) |
| Injury grade | |
| I | 4 (19.0) |
| II | 7 (33.3) |
| III | 5 (23.8) |
| IV | 4 (19.0) |
| V | 1 (4.8) |
Values are shown as the number of patients (%) unless indicated otherwise.
There were 4 deaths (19.0%) in the BCVI cohort. Additionally, 14 (66.7%) of these patients had one or more intracranial injuries, including 1 epidural hematoma, 7 subdural hematomas, 11 subarachnoid hemorrhages (with 1 patient later developing vasospasm), and 4 contusions (Table 4). Six (28.6%) patients had a stroke: 2 patients had a stroke on initial presentation and 4 patients had a stroke within 24–96 hours after initial medical management. These 4 patients had late strokes despite medical management with antiplatelet therapy. Furthermore, the majority of the BCVI patients (16; 76.2%) sustained a carotid artery injury (9 [56.3%] were extracranial carotid artery injuries and 7 [43.8%] were intracranial carotid artery injuries), while the remaining 5 patients (23.8%) sustained a vertebral artery injury (all 5 [100%] were extracranial vertebral artery injuries, as shown in Table 4). Of these arterial injuries, 4 injuries (19.0%) were grade I, 7 injuries (33.3%) were grade II, 5 injuries (23.8%) were grade III, 4 injuries (19.0%) were grade IV, and 1 injury (4.8%) was grade V according to the Biffl scale7 (Table 4). Additionally, a total of 13 patients (61.9%) received treatment, while the remaining 8 patients (38.1%) were observed and received no medication (Table 5). Of the patients with BCVI who were treated (n = 13), 2 patients (15.4%) had grade I injuries, 5 patients (38.5%) had grade II injuries, 3 patients (23.1%) had grade III injuries, and 3 patients (23.1%) had grade IV injuries (Table 4). Twelve of the 13 patients who received medical management were treated with antiplatelet therapy while 1 was treated with anticoagulation. However, this patient was started on anticoagulation therapy to treat deep venous sinus thrombosis and not for BCVI.
Patient and injury characteristics of the patients with and without BCVI
| Variable | No BCVI | BCVI | p Value* |
|---|---|---|---|
| No. of patients | 439 | 21 | |
| MOI, n (%) | |||
| Motor vehicle accident | 209 (47.6) | 11 (52.4) | 0.67 |
| Automobile-pedestrian accident | 37 (8.4) | 5 (23.8) | 0.02† |
| Fall | 64 (14.6) | 2 (9.5) | 0.51 |
| Bicycle accident | 12 (2.7) | 1 (4.8) | 0.57 |
| Other | 117 (26.7) | 2 (9.5) | 0.08 |
p values were determined using the chi-square test.
Significant at α = 0.05.
Summary of the treatment received by each patient in the cohort of patients diagnosed with BCVI (n = 21) for each grade of vascular injury
| Treatment | Grade of Vascular Injury* | ||||
|---|---|---|---|---|---|
| Grade I | Grade II | Grade III | Grade IV | Grade V | |
| Observation | 2 | 2 | 2 | 1 | 1 |
| Antiplatelet therapy | 1 | 5 | 3 | 3 | 0 |
| Anticoagulant therapy | 1† | 0 | 0 | 0 | 0 |
| Endovascular | 0 | 0 | 0 | 0 | 0 |
| Open surgery | 0 | 0 | 0 | 0 | 0 |
Values are presented as the number of patients.
The grade of vascular injury was determined using the Biffl grading scale.
This patient was managed with anticoagulation therapy for deep venous sinus thrombosis and not BCVI.
Utah Score Variables
The distribution of the Utah scores in patients who underwent angiography and were not diagnosed with BCVI (n = 439) was significantly (p < 0.05) different from the distribution of Utah scores in the cohort of patients diagnosed with BCVI (n = 21). In the imaging cohort, the clear majority—275 patients (62.6%)—had a Utah score of 0; 104 (23.7%), 22 (5.0%), 22 (5.0%), 10 (2.3%), 3 (0.7%), and 3 (0.7%) patients had Utah scores of 1, 2, 3, 4, 5, and 6, respectively. No patient (0%) had a Utah score greater than 6. In the group of patients diagnosed with BCVI, 4 (19.0%), 5 (23.8%), 1 (4.8%), 3 (14.3%), 2 (9.5%), 2 (9.5%), 1 (4.8%), 2 (9.5%), and 1 (4.8%) patients had Utah scores of 0, 1, 2, 3, 4, 5, 6, 7, and 8, respectively (Table 6).
Comparison of Utah scores between the BCVI and no-BCVI cohorts
| Utah Score | No. of Patients (%) | p Value* | |
|---|---|---|---|
| No-BCVI Cohort (n = 439) | BCVI Cohort (n = 21) | ||
| 0 | 275 (62.6) | 4 (19.0) | <0.001† |
| 1 | 104 (23.7) | 5 (23.8) | 0.99 |
| 2 | 22 (5.0) | 1 (4.8) | 0.97 |
| 3 | 22 (5.0) | 3 (14.3) | 0.06 |
| 4 | 10 (2.3) | 2 (9.5) | 0.04† |
| 5 | 3 (0.7) | 2 (9.5) | <0.001† |
| 6 | 3 (0.7) | 1 (4.8) | 0.05 |
| 7 | 0 (0) | 2 (9.5) | <0.001† |
| 8 | 0 (0) | 1 (4.8) | <0.001† |
Determined using the chi-square test.
Significant at α = 0.05.
Specifically, among the group of patients who underwent angiography and were not found to have BCVI (n = 439), 133 patients (30.3%) had GCS score ≤ 8, 21 patients (4.8%) had a focal neurological deficit, 20 patients (4.6%) had a fracture through the carotid canal, 15 patients (3.4%) had a petrous temporal bone fracture, and 7 patients (1.6%) had hypodensity on CT that was consistent with ischemia (Table 7). These values contrasted significantly (p < 0.05) from those seen among patients diagnosed with BCVI (n = 21) in which 10 patients (47.6%) diagnosed had GCS score ≤ 8, 4 patients (19.0%) had a focal neurological deficit, 8 patients (38.1%) had a carotid canal fracture, 6 patients (28.6%) had a petrous temporal bone fracture, and 2 patients (9.5%) had a cerebral infarction on CT at initial presentation (Table 7).
Comparison of the patients who had at least 1 of the 5 Utah variables in the BCVI and no-BCVI cohorts
| Utah Variable | No. of Patients (%) | p Value* | |
|---|---|---|---|
| No-BCVI Cohort | BCVI Cohort | ||
| No. of patients | 439 | 21 | |
| GCS score ≤8 | 133 (30.3) | 10 (47.6) | 0.09 |
| Focal neurological deficit | 21 (4.8) | 4 (19.0) | 0.01† |
| Carotid canal fracture | 20 (4.6) | 8 (38.1) | <0.001† |
| Petrous temporal bone fracture | 15 (3.4) | 6 (28.6) | <0.001† |
| Cerebral infarction | 7 (1.6) | 2 (9.5) | 0.01† |
Determined using the chi-square test.
Significant at α = 0.05.
Assessment of the Utah Score
Each patient who received angiographic imaging (n = 460) was scored using the Utah scoring system and classified as high risk (score ≥ 3) or low risk (score ≤ 2). Using this threshold for high and low risk, a sensitivity of 52.4%, specificity of 91.3%, positive predictive value of 22.4%, and negative predictive value of 97.6% were calculated (Table 2). In total, 47.6% (10 of 21) of patients with BCVI were misclassified as low risk (false negative) and 8.7% (38 of 439) of patients without BCVI were misclassified as high risk (false positive).
To develop a more sensitive screening tool for identifying patients at high risk for BCVI, a linear regression model was generated to determine the most sensitive and specific Utah score. A Utah score ≥ 1 produced a greater AUC (78.9%) than a score ≥ 3 (72%). A Utah score ≥ 1 was redefined as high risk, and patients with a Utah score of 0 were redefined as low risk. With a Utah score ≥ 1, 19.0% (4 of 21) of BCVI patients were misclassified as low risk (false negative) and 37.4% (164 of 439) of patients without BCVI were misclassified as high risk (false positive), as shown in Table 2.
To reduce the number of false positives associated with a Utah score ≥ 1, the MOI was incorporated into the criteria for defining high versus low risk. The MOI was considered because of the greater number of BCVI patients involved in a motor vehicle collision or automobile-pedestrian accident than patients without BCVI. No other combination of demographic or clinical variables collected by our institution or measured by other screening scores (Denver, modified Memphis, or EAST) led to improved sensitivity and specificity within our cohort. The new screening score, the McGovern score, defined high risk as a McGovern score ≥ 3. The high-risk McGovern score was significantly associated with BCVI with an odds ratio of 10.6 (95% CI 3.5–32.0; p < 0.0001). The score discriminated well, with an AUC of 83.4% (Fig. 1). Furthermore, the increase in AUC when incorporated with the MOI into the McGovern score was significant (p < 0.05) compared with a Utah score ≥ 1. When the McGovern score was applied to the BCVI cohort, a sensitivity of 81.0%, specificity of 71.3%, positive predictive value of 11.9%, and negative predictive value of 98.7% were calculated (Table 2).
Assessment of Other Screening Scores
The criteria for the Denver,18 modified Memphis,16 and EAST12 screening tools are summarized in Table 8. In our cohort, there were no neck soft-tissue injuries, no Le Fort fractures, and no cases of Horner’s syndrome. There were, however, 3 patients with cervical spine involvement (2 fractures and 1 subluxation) and 2 instances of anoxic brain injury on initial presentation. Of the 21 cases of BCVI in our cohort, the Denver criteria would have missed 6 cases (sensitivity of 71%), the modified Memphis criteria would have missed 6 cases (sensitivity of 71%), and the EAST criteria would have missed 7 cases (sensitivity of 67%). None of these scores incorporated the MOI into their screening criteria. Furthermore, no BCVI patient in our cohort (n = 21) was correctly identified using the other screening scores (Denver, modified Memphis, EAST, and Utah) but missed using the McGovern score.
Summaries of the Denver, modified Memphis, and EAST screening criteria for BCVI
| Denver Criteria | Modified Memphis Criteria | EAST Criteria |
|---|---|---|
| Focal neurological deficit | Petrous temporal bone fracture | Cervical hyperextension associated w/ displaced midface or complex mandibular fracture or closed head injury consistent with diffuse axonal injury |
| Arterial hemorrhage | Carotid canal fracture | Anoxic brain injury due to hypoxia as a result of squeezed arteries |
| Cervical bruit in patients <50 yrs | Le Fort fracture II or III | Seatbelt abrasion or other soft-tissue injury resulting in swelling or altered mental status |
| Expanding neck hematoma | Cervical spine fracture | Cervical vertebral body fracture or carotid canal fracture in proximity to the internal carotid or vertebral arteries |
| Neurological exam findings inconsistent w/ head CT scan | Horner’s syndrome | |
| Cerebrovascular accident on follow-up head CT scan not seen on initial head CT scan | Neck soft-tissue injury (seatbelt sign, hypoxia as a result of squeezed arteries, or hematoma) | |
| Presence of Le Fort II or III fractures | Focal neurologic deficit not explained by imaging | |
| Cervical spine fracture w/ subluxation | ||
| C1–3 cervical spine fracture | ||
| Cervical spine fracture extending into the transverse foramen | ||
| Basilar skull fracture w/ carotid involvement | ||
| Diffuse axonal injury w/ GCS score <6 | ||
| Hypoxic ischemia due to squeezed arteries |
For each of these 3 screening tools, if any of the screening criteria are met, the recommendation is to perform further workup with angiographic imaging.
Discussion
In this study, we reviewed a prospectively collected cohort of 12,614 pediatric patients with blunt trauma and identified 460 patients (3.6%) who received angiographic imaging of the head and/or neck. Of those screened, 21 patients (4.6%) were diagnosed with BCVI.
Predicting BCVI in Children
There has been a recent flurry of research regarding BCVI in children,8,21,37,41,42 with a special emphasis on the development and validation of screening criteria. While previous screening criteria have been used to evaluate the need for angiographic imaging in adult patients at high risk for BCVI (Denver and modified Memphis screening criteria), the Utah score has been proposed as a potential screening system for the pediatric trauma population. The Utah score assigns points based on the presence or absence of 5 BCVI risk factors with a maximum score of 11: GCS score ≤ 8 (1 point), focal neurological deficit (2 points), cerebral infarction on CT (3 points), fracture through the carotid canal (2 points), and petrous temporal bone fracture (3 points). The creators of the Utah score assigned a value of ≤ 2 as low risk, which corresponds to a posttest probability of 7.9%. In a subsequent multicenter validation study, the same group found a total misclassification rate of 16.6% among patients who underwent CTA at 4 Level 1 pediatric trauma centers.
In our study, we retrospectively applied the Utah, modified Memphis, Denver, and EAST screening criteria to 21 patients who were diagnosed with BCVI. In doing so, we found that the modified Memphis, Denver, and EAST screening criteria misclassified 6 (28.6%), 6 (28.6%), and 7 (33.3%) patients with BCVI, respectively, as low risk and not in need of subsequent angiographic imaging. Furthermore, the Utah score misclassified 10 patients (47.6%) with BCVI as low risk, including 5 patients with an injury score ≥ 3 according to the Biffl scale,7 which is used to classify the severity of BCVI (grade I is mild intimal injury, grade II is dissection or luminal narrowing, grade III is pseudoaneurysm, grade IV is vessel occlusion, and grade V is vessel transection). This finding is consistent with a recent multicenter validation study of the Utah score in which 9 of 22 patients with BCVI were misclassified as low risk (40.9%).41 The authors of that study correctly noted that the negative predictive value of the Utah score is quite high (97%), but this must be taken in the context of the rarity of BCVI in children. In our study, less than 1% of all pediatric patients with blunt trauma and only 4.6% of patients who received angiographic imaging had BCVI. Thus, hypothetical screening criteria that misclassify all true BCVI patients as low risk would have a negative predictive value between 95% and 100%.
One of our primary motivations for developing a better screening tool for determining BCVI risk is to better standardize the assessment of trauma patients between academic centers. Currently, there are no standards by which institutions, including our own, perform angiography or treat pediatric BCVI. The decision to obtain images is largely based on the severity of the injury and clinical gestalt. Without a reliable method by which institutions obtain angiographic imaging to consistently identify pediatric BCVI, a standardized treatment protocol (e.g., observation, anticoagulation, and surgical management) cannot be developed because the outcomes between institutions will not be comparable and many BCVI patients will not be identified. Furthermore, the incidence of pediatric BCVI is debatable, with some studies suggesting that it may be as high as 9%.24 To more accurately determine the incidence and applicability of BCVI treatment, a good screening tool is a necessity. Ultimately, an accurate pediatric BCVI screening tool allows for a standardized BCVI assessment across academic institutions and provides a necessary platform for future research to be built on.
Another important consideration in the development of a better screening tool for determining BCVI risk is to prevent pediatric patients from being exposed to unnecessary radiation. Longitudinal studies have demonstrated that patients exposed to radiation in childhood have an increased risk for cancer development. Specifically, children younger than 10 years who underwent CTA, which corresponds to about 12–16 mSv of radiation, have a 0.18%–0.24% increased cancer risk above baseline per angiography study.2,31 Furthermore, nearly one-quarter of our cohort underwent multiple angiography sessions. Based on this information, obtaining angiograms in every patient in our cohort (n = 460) suspected of having BCVI would cause secondary cancers in 1–2 patients due to radiation exposure. This long-term cancer risk is concerning because the McGovern score misclassified 4 patients diagnosed with BCVI, and all 4 of these patients were treated conservatively (3 with observation and 1 with aspirin). For this reason, the risk of cancer development outweighs the benefits of broad angiography. On the other hand, while obtaining angiograms in every patient is unacceptable, efforts should be taken to develop screening criteria that will misclassify as few patients as possible. As newer radiation techniques and more optimal angiography modalities continue to be developed, we suspect that the radiation exposure that results from these studies will decrease over time; however, due to concerns regarding the long-term risk of radiation exposure in pediatric patients, we feel than an improved screening tool will still benefit this patient population by reducing unnecessary radiation exposure.
Synthesizing Clinical Findings and MOI: The McGovern Score
Despite the high rate of misclassification, the Utah score nonetheless provides a great starting point for the creation of a clinically useful tool for pediatric BCVI screening. As opposed to all other aforementioned screening scales (Denver, modified Memphis, and EAST), the Utah score was specifically designed for a pediatric trauma cohort and had been previously validated at another institution.41 In our series, we found that many of the patients with high-grade BCVI who were misclassified shared in common a high-impact MOI involving motor vehicles (the patient was either a passenger in a car or motorcycle accident or a pedestrian struck by a car or motorcycle). For example, of the 10 patients who were subsequently found to have BCVI on angiography and were misclassified as low risk by the Utah score, 9 were involved in a motor vehicle accident. Furthermore, the only risk factors for BCVI in the majority of patients misclassified by the Denver, modified Memphis, or EAST screening criteria were GCS score ≤ 8 and being involved in a motor vehicle accident. There were no other significant clinical or radiological findings consistently demonstrated in our BCVI cohort (n = 21). It would be unreasonable, however, to recommend angiographic studies for every case of pediatric trauma resulting from motor vehicle accidents because this would result in a huge number of unnecessary studies given the low incidence of pediatric BCVI.
We propose modifying the Utah score by assigning the MOI a point value of 2. Thus, patients who present with either a Utah score ≥ 3 or a Utah score ≥ 1 plus the MOI will be screened. The McGovern score was 28.6% more sensitive for predicting pediatric BCVI than a Utah score ≥ 3, 14.3% more sensitive for predicting pediatric BCVI than the EAST screening criteria, 9.5% more sensitive for predicting pediatric BCVI than the modified Memphis screening criteria, and 9.5% more sensitive for predicting pediatric BCVI than the Denver screening criteria. By including the presence of a high-speed traumatic mechanism due to a motor vehicle accident, the McGovern score captures and expands on the seatbelt criteria included in the EAST and modified Memphis criteria to include those patients involved in a motor vehicle accident but without the seatbelt sign. Most importantly, the McGovern score excluded none of the high-grade injuries found in our cohort.
One of the major difficulties with identifying BCVI is that the majority of patients exhibit no focal neurological deficits on initial presentation to the emergency department.6 In the adult population, neurological symptoms most often occur 10–72 hours after injury,17,19 and studies have shown that this delay in the presentation of symptoms also occurs in the pediatric BCVI population.28 While the Denver, modified Memphis, and EAST screening criteria excelled at identifying patients with high-grade vascular injuries in our cohort, they were unable to identify the group of patients with BCVI who did not present with prominent signs of acute injury. In our BCVI cohort (n = 21), 7 patients (33%) on initial presentation had no focal neurological deficits, fractures, or soft-tissue signs that could be identified on initial imaging. Furthermore, 4 of those 7 patients had GCS > 9. Using the data collected, no combination of criteria mentioned in the previous studies would have correctly identified these missed patients without significantly lowering the specificity of our screening tool. Of those 7 patients, the Denver criteria missed 6 patients, the modified Memphis criteria missed 6 patients, the EAST screening criteria missed all 7 patients, and the Utah score missed all 7 patients. The McGovern score, however, only missed 4 of these patients and correctly identified 2 patients who had embolic strokes found on head CT at 7 days and 3 days, respectively, after initial presentation and were missed by all of the other screening scores. In total, 6 patients were found to have strokes: 2 patients on admission and 4 patients within 24–96 hours after admission despite medical management with antiplatelet therapy. The McGovern score identified all 6 patients, while the Utah score, Denver criteria, modified Memphis criteria, and EAST criteria identified 3, 2, 2, and 2 patients, respectively. Furthermore, none of the 4 patients missed by the McGovern score were correctly identified by the Denver, modified Memphis, EAST, or Utah scores.
The goal of a good screening tool is to identify the patients with BCVI who do not present with prominent signs of deficit and do not have clear risk factors on imaging. Previous studies have determined that at least 30% of BCVI patients will not meet conventional screening criteria,13 and 29%–33% of patients fell through the cracks of conventional screening criteria (Denver, modified Memphis, and EAST) in our cohort. Numerous studies that investigated BCVI have explored the effects that different MOIs have on BCVI incidence, with nearly all of them reporting that most BCVI patients were involved in a motor vehicle accident.8,22,30,35,41 However, to our knowledge, this is the first study to incorporate the MOI into a screening tool for predicting BCVI.
Looking to the Future: Determining the Optimal Treatment for Pediatric BCVI
As in previous studies,21 we found that many of our patients were treated with observation alone (38.1%). None of our BCVI patients (n = 21) were managed with surgical or endovascular intervention. Of the 13 patients (61.9%) with BCVI who received medical management, 11 patients (84.6%) were treated with aspirin, 1 patient (7.7%) was treated with aspirin and Plavix, and 1 patient (7.7%) was being treated with anticoagulation (Lovenox); however, this treatment was for concurrent deep venous sinus thrombosis and not BCVI management. None of the 4 patients in our cohort who died (19%) would likely have benefitted from intervention for their BCVI: 2 patients were found to have a global hypoxic-ischemic injury caused by cerebral vascular damage as a result of the trauma that was present on admission and deemed nonsurvivable; the other 2 patients presented with grade III diffuse axonal injury and absent brainstem reflexes on admission. As reported by Dewan et al.,21 the main prognostic factor in our cohort appears to be noncerebrovascular injuries present on admission. Also of note is that 11 of 21 patients had subarachnoid hemorrhage on initial presentation, but only 1 of those patients (1 of 21 patients; 4.8%) later developed vasospasm, which is substantially lower than the vasospasm incidence of 10.6% in the adult BCVI population14 versus approximately 40% in adults with severe head injuries.48 A noninvasive, nonradiation alternative is to use transcranial Doppler and careful observation to assess the possible future development of vasospasm in patients who present with subarachnoid hemorrhage.
There is no consensus regarding the optimal treatment strategies for pediatric BCVI. Our patients were treated conservatively, and none of them had deficits attributable to their vascular injury on follow-up. However, we appreciate that as techniques for clot removal and stenting advance, more children with severe vascular injuries on CTA will potentially receive endovascular treatment in the future. Additionally, it is important to note that pediatric BCVI is substantially less severe than adult BCVI at the same center as reported by our group.38 We suspect that this lack of severity is multifactorial. Children have a relatively greater elastic resilience of their vessels compared with adults and have more elastic bone and soft tissues surrounding the vessels that can better absorb the kinetic energy of high-impact blunt trauma. There was no craniocaudal anatomical injury predilection in our cohort. In addition, children have less-diseased vessels, which allows for better recovery even in the setting of injury. In line with the previously referenced studies, our analysis suggests that many pediatric patients diagnosed with BCVI by using angiographic imaging may be safely managed conservatively, i.e., with observation or antiplatelet therapy alone. More research is needed to identify the optimal treatment strategy for pediatric BCVI, and a standardized screening tool for identifying BCVI by institutions will be useful for the development of appropriate BCVI management. Moving forward, we are prospectively collecting McGovern screening scores at our institution to validate the tool’s utility. Additionally, we are looking to validate this score with multicenter data and have already requested these data from the Utah group. By applying our screening score to data collected from other institutions, we intend to determine if our screening score does a better job of identifying high-risk BCVI patients.
Limitations
Due to the high-volume of pediatric patients with trauma at our institution, we were able to conduct a single-institution study with a cohort of patients that is as large as those included in other multicenter studies. In this way, we hoped to provide important, independent data to assess the generalizability of other BCVI screening criteria (modified Memphis, Denver, EAST, and Utah). However, because this is a single-center retrospective cohort study, it suffers from inherent institutional bias in terms of screening and treatment paradigms. Our center does not currently have a protocol for treating pediatric BCVI, as reflected by the wide variety of interventions, or lack thereof, in our patient cohort. Given that several of our patients who had BCVI had few neurological signs on admission (especially those with low-grade injury), it is reasonable to assume that our incidence of BCVI would have been higher if all of the patients in the initial cohort (12,614 in total) had undergone angiographic screening. Because obtaining angiograms in all 12,614 patients would be impractical, it is impossible to determine the exact incidence of BCVI in our cohort. Additionally, the radiation involved in such an aggressive screening strategy would likely be harmful to the pediatric population. Over the time period of this study, our institution used both CTA and MRA as diagnostic tests to confirm BCVI, although we exclusively use CTA now. Research regarding the differences in the accuracies of CTA and MRA for diagnosing BCVI has yielded mixed results.46 It is possible, but unlikely—because no significant regions of diffusion restriction were seen—that clinically significant cases of BCVI may have been missed in our cohort of 107 patients (23%) who only underwent MRA with concomitant MRI and not CTA or angiography.
Conclusions
This study reviewed a prospectively collected Level 1 pediatric trauma registry and developed a novel grading system, the McGovern score, to help predict the presence of BCVI from the initial head CT and the MOI to help avoid unnecessary radiation with CTA. The McGovern score had a sensitivity of 81.0%, specificity of 71.3%, and negative predictive value of 98.7% for predicting BCVI, and a specificity of 100% for predicting eventual associated stroke (28.6%). The McGovern score was superior to the Denver, East, modified Memphis, and Utah criteria for predicting BCVI as well as eventual associated stroke and provides a more reliable and standardized assessment of BCVI across academic institutions. Furthermore, because no level of ionizing radiation has been proven safe in children, prospective validation and utilization of the McGovern score could be valuable for reducing the long-term iatrogenic cancer risk in traumatically injured children.
Acknowledgments
We acknowledge the support provided by the Biostatistics/Epidemiology/Research Design component of the Center for Clinical and Translational Sciences (CCTS) for this project. CCTS is mainly funded by a grant (UL1 TR000371) from the National Center for Advancing Translational Sciences (NCATS) that was awarded to the University of Texas Health Science Center at Houston. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NCATS. The research reported in this study was supported by the Eunice Kennedy Shriver National Institute of Child Health & Human Development of the National Institutes of Health under award no. L30 HD089125 (awarded to M.N.S.).
Disclosures
The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper.
Author Contributions
Conception and design: Shah. Acquisition of data: Shah, Herbert, Venkataraman, Turkmani, Kerr. Analysis and interpretation of data: Shah, Herbert, Venkataraman, Turkmani, Zhu, Patel, Ugalde, Fletcher, Sandberg, Cox, Kitagawa, Day. Drafting the article: Shah, Herbert, Venkataraman. Critically revising the article: all authors. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Shah. Statistical analysis: Herbert, Venkataraman, Zhu. Study supervision: Shah.
Supplemental Information
Previous Presentations
Portions of this work were presented as an abstract at the 45th Annual Meeting of the AANS/CNS Section on Pediatric Neurosurgery, Orlando, FL, December 5–8, 2016.
References
- 1↑
Abdel-Aziz H, Dunham CM, Malik RJ, Hileman BM: Timing for deep vein thrombosis chemoprophylaxis in traumatic brain injury: an evidence-based review. Crit Care 19:96, 2015
- 2↑
Alessio AM, Phillips GS: A pediatric CT dose and risk estimator. Pediatr Radiol 40:1816–1821, 2010
- 3↑
Anaya C, Munera F, Bloomer CW, Danton GH, Caban K: Screening multidetector computed tomography angiography in the evaluation on blunt neck injuries: an evidence-based approach. Semin Ultrasound CT MR 30:205–214, 2009
- 5↑
Azarakhsh N, Grimes S, Notrica DM, Raines A, Garcia NM, Tuggle DW, et al.: Blunt cerebrovascular injury in children: underreported or underrecognized? A multicenter ATOMAC study. J Trauma Acute Care Surg 75:1006–1012, 2013
- 6↑
Bachier M, Eubanks JW III: Blunt cerebrovascular injury in children: when is aggressive management necessary? Curr Surg Rep 2:64, 2014
- 7↑
Biffl WL, Moore EE, Offner PJ, Brega KE, Franciose RJ, Burch JM: Blunt carotid arterial injuries: implications of a new grading scale. J Trauma 47:845–853, 1999
- 8↑
Biffl WL, Moore EE, Offner PJ, Brega KE, Franciose RJ, Elliott JP, et al.: Optimizing screening for blunt cerebrovascular injuries. Am J Surg 178:517–522, 1999
- 9↑
Biffl WL, Moore EE, Ryu RK, Offner PJ, Novak Z, Coldwell DM, et al.: The unrecognized epidemic of blunt carotid arterial injuries: early diagnosis improves neurologic outcome. Ann Surg 228:462–470, 1998
- 10↑
Bonatti M, Vezzali N, Ferro F, Manfredi R, Oberhofer N, Bonatti G: Blunt cerebrovascular injury: diagnosis at whole-body MDCT for multi-trauma. Insights Imaging 4:347–355, 2013
- 11↑
Brenner DJ, Hall EJ: Computed tomography—an increasing source of radiation exposure. N Engl J Med 357:2277–2284, 2007
- 12↑
Bromberg WJ, Collier BC, Diebel LN, Dwyer KM, Holevar MR, Jacobs DG, et al.: Blunt cerebrovascular injury practice management guidelines: the Eastern Association for the Surgery of Trauma. J Trauma 68:471–477, 2010
- 13↑
Bruns BR, Tesoriero R, Kufera J, Sliker C, Laser A, Scalea TM, et al.: Blunt cerebrovascular injury screening guidelines: what are we willing to miss? J Trauma Acute Care Surg 76:691–695, 2014
- 14↑
Buch K, Nguyen T, Mahoney E, Libby B, Calner P, Burke P, et al.: Association between cervical spine and skull-base fractures and blunt cerebrovascular injury. Eur Radiol 26:524–531, 2016
- 15↑
Burlew CC, Biffl WL, Moore EE, Pieracci FM, Beauchamp KM, Stovall R, et al.: Endovascular stenting is rarely necessary for the management of blunt cerebrovascular injuries. J Am Coll Surg 218:1012–1017, 2014
- 16↑
Ciapetti M, Circelli A, Zagli G, Migliaccio ML, Spina R, Alessi A, et al.: Diagnosis of carotid arterial injury in major trauma using a modification of Memphis criteria. Scand J Trauma Resusc Emerg Med 18:61, 2010
- 17↑
Cothren CC, Biffl WL, Moore EE, Kashuk JL, Johnson JL: Treatment for blunt cerebrovascular injuries: equivalence of anticoagulation and antiplatelet agents. Arch Surg 144:685–690, 2009
- 19↑
Cothren CC, Moore EE, Biffl WL, Ciesla DJ, Ray CE Jr, Johnson JL, et al.: Anticoagulation is the gold standard therapy for blunt carotid injuries to reduce stroke rate. Arch Surg 139:540–546, 2004
- 20↑
Desai NK, Kang J, Chokshi FH: Screening CT angiography for pediatric blunt cerebrovascular injury with emphasis on the cervical “seatbelt sign”. AJNR Am J Neuroradiol 35:1836–1840, 2014
- 21↑
Dewan MC, Ravindra VM, Gannon S, Prather CT, Yang GL, Jordan LC, et al.: Treatment practices and outcomes after blunt cerebrovascular injury in children. Neurosurgery 79:872–878, 2016
- 22↑
DiCocco JM, Fabian TC, Emmett KP, Magnotti LJ, Zarzaur BL, Bate BG, et al.: Optimal outcomes for patients with blunt cerebrovascular injury (BCVI): tailoring treatment to the lesion. J Am Coll Surg 212:549–559, 2011
- 23↑
Einstein AJ, Henzlova MJ, Rajagopalan S: Estimating risk of cancer associated with radiation exposure from 64-slice computed tomography coronary angiography. JAMA 298:317–323, 2007
- 24↑
Esnault P, Cardinale M, Boret H, D’Aranda E, Montcriol A, Bordes J, et al.: Blunt cerebrovascular injuries in severe traumatic brain injury: incidence, risk factors, and evolution. J Neurosurg 127:16–22, 2017
- 25↑
Fabian TC, Patton JH Jr, Croce MA, Minard G, Kudsk KA, Pritchard FE: Blunt carotid injury. Importance of early diagnosis and anticoagulant therapy. Ann Surg 223:513–525, 1996
- 26↑
Fleck SK, Langner S, Baldauf J, Kirsch M, Kohlmann T, Schroeder HWS: Incidence of blunt craniocervical artery injuries: use of whole-body computed tomography trauma imaging with adapted computed tomography angiography. Neurosurgery 69:615–624, 2011
- 27↑
Franz RW, Willette PA, Wood MJ, Wright ML, Hartman JF: A systematic review and meta-analysis of diagnostic screening criteria for blunt cerebrovascular injuries. J Am Coll Surg 214:313–327, 2012
- 28↑
Fullerton HJ, Johnston SC, Smith WS: Arterial dissection and stroke in children. Neurology 57:1155–1160, 2001
- 29↑
Hausleiter J, Meyer T, Hermann F, Hadamitzky M, Krebs M, Gerber TC, et al.: Estimated radiation dose associated with cardiac CT angiography. JAMA 301:500–507, 2009
- 30↑
Hills NK, Johnston SC, Sidney S, Zielinski BA, Fullerton HJ: Recent trauma and acute infection as risk factors for childhood arterial ischemic stroke. Ann Neurol 72:850–858, 2012
- 31↑
Ivanov VK, Kashcheev VV, Chekin SY, Menyaylo AN, Pryakhin EA, Tsyb AF, et al.: Estimation of risk from medical radiation exposure based on effective and organ dose: how much difference is there? Radiat Prot Dosimetry 155:317–328, 2013
- 32↑
Jones TS, Burlew CC, Kornblith LZ, Biffl WL, Partrick DA, Johnson JL, et al.: Blunt cerebrovascular injuries in the child. Am J Surg 204:7–10, 2012
- 33↑
Kerwin AJ, Bynoe RP, Murray J, Hudson ER, Close TP, Gifford RR, et al.: Liberalized screening for blunt carotid and vertebral artery injuries is justified. J Trauma 51:308–314, 2001
- 34↑
Kleinerman RA: Cancer risks following diagnostic and therapeutic radiation exposure in children. Pediatr Radiol 36 (Suppl 2):121–125, 2006
- 35↑
Kopelman TR, Berardoni NE, O’Neill PJ, Hedayati P, Vail SJ, Pieri PG, et al.: Risk factors for blunt cerebrovascular injury in children: do they mimic those seen in adults? J Trauma 71:559–564, 2011
- 36↑
Li W, D’Ayala M, Hirshberg A, Briggs W, Wise L, Tortolani A: Comparison of conservative and operative treatment for blunt carotid injuries: analysis of the National Trauma Data Bank. J Vasc Surg 51:593–599, 599.e1–599.e2, 2010
- 37↑
Malhotra A, Wu X, Kalra VB, Goodman TR, Schindler J, Forman HP: Screening for pediatric blunt cerebrovascular injury: review of literature and a cost-effectiveness analysis. J Pediatr Surg 50:1751–1757, 2015
- 38↑
McNutt MK, Kale AC, Kitagawa RS, Turkmani AH, Fields DW, Baraniuk S, Gill BS, Cotton BA, Moore LJ, Wade CE, Day A, Holcomb JB: Management of blunt cerebrobascular injury (BCVI) in the multisystem injury patient with contraindications to immediate anti-thrombotic therapy. Injury 49:67–74, 2018
- 39↑
Miglioretti DL, Johnson E, Williams A, Greenlee RT, Weinmann S, Solberg LI, et al.: The use of computed tomography in pediatrics and the associated radiation exposure and estimated cancer risk. JAMA Pediatr 167:700–707, 2013
- 40↑
Pearce MS, Salotti JA, Little MP, McHugh K, Lee C, Kim KP, et al.: Radiation exposure from CT scans in childhood and subsequent risk of leukaemia and brain tumours: a retrospective cohort study. Lancet 380:499–505, 2012
- 41↑
Ravindra VM, Bollo RJ, Sivakumar W, Akbari H, Naftel RP, Limbrick DD Jr, et al.: Predicting blunt cerebrovascular injury in pediatric trauma: validation of the “Utah Score”. J Neurotrauma 34:391–399, 2017
- 42↑
Ravindra VM, Riva-Cambrin J, Sivakumar W, Metzger RR, Bollo RJ: Risk factors for traumatic blunt cerebrovascular injury diagnosed by computed tomography angiography in the pediatric population: a retrospective cohort study. J Neurosurg Pediatr 15:599–606, 2015
- 43↑
Sarkar K, Keachie K, Nguyen U, Muizelaar JP, Zwienenberg-Lee M, Shahlaie K: Computed tomography characteristics in pediatric versus adult traumatic brain injury. J Neurosurg Pediatr 13:307–314, 2014
- 44↑
Stein DM, Boswell S, Sliker CW, Lui FY, Scalea TM: Blunt cerebrovascular injuries: does treatment always matter? J Trauma 66:132–144, 2009
- 45↑
Tso MK, Lee MM, Ball CG, Morrish WF, Mitha AP, Kirkpatrick AW, et al.: Clinical utility of a screening protocol for blunt cerebrovascular injury using computed tomography angiography. J Neurosurg 126:1033–1041, 2017
- 46↑
Vertinsky AT, Schwartz NE, Fischbein NJ, Rosenberg J, Albers GW, Zaharchuk G: Comparison of multidetector CT angiography and MR imaging of cervical artery dissection. AJNR Am J Neuroradiol 29:1753–1760, 2008
- 47↑
Wei CW, Montanera W, Selchen D, Lian J, Stevens C, de Tilly LN: Blunt cerebrovascular injuries: diagnosis and management outcomes. Can J Neurol Sci 37:574–579, 2010
- 48↑
Zurynski YA, Dorsch NW: A review of cerebral vasospasm. Part IV. Post-traumatic vasospasm. J Clin Neurosci 5:146–154, 1998
