Are Skull Radiographs Indicated in Pediatric Head Trauma?
Are Skull Radiographs Indicated in Pediatric Head Trauma?
Author: Samir Haydar, DO, MPH, Clinical Instructor, Department of Emergency Medicine, Maine Medical Center, Portland.
Peer Reviewer: Milan D. Nadkarni, MD, Director, Pediatric Emergency Department, Wake Forest Baptist Medical Center, Associate Professor, Pediatrics and Emergency Medicine, Wake Forest University, Winston Salem, NC.
Introduction
Pediatric head injury is the leading cause of trauma mortality in children. The adult literature supports the use of clinical criteria (such as loss of consciousness, Glasgow Comma Scale [GCS] <15, worsening headache, clinical evidence of basilar skull fracture, multiple episodes vomiting, and evidence of focal neurologic defects) to focus diagnostic testing.1,2 However, injury severity often is more elusive in the pediatric population. As a result, the clinical signs and symptoms present in young children have not been shown to serve as sensitive predictors of injury.
In addition, due to its increased availability and overall sensitivity, computed tomography (CT) has come to the forefront of this diagnostic paradigm. Although medical-legal issues likely play a role in this trend, an increasing concern for high levels of central nervous system radiation exposure with CT has reopened the controversy of alternative imaging techniques such as skull radiographs. Due primarily to their lack of sensitivity, skull radiographs are thought to have limited utility in the management of pediatric head injuries. Central to this debate is the understanding that skull fractures may occur without damage to the brain and, conversely, that severe brain injury can occur in the absence of skull fractures. Adding skull radiographs to protocols intended to risk stratify head injuries may effectively reduce our reliance on CT scans without resulting in missed injuries. Unfortunately, the current (past 2-3 years) pediatric literature database is essentially devoid of clinical trials involving the use of skull radiographs as a diagnostic option for head injured patients. Therefore, the clinical utility of such must be inferred from studies completed before radiation exposure from CT was considered.
Is There a Role for Skull X-rays in Head Injury?
Source: Reed MJ, Browning JG, Wilkinson AG, et al. Can we abolish skull x rays for head injury? Arch Dis Child 2005;90:859-864.
The use of skull radiographs was fairly well studied prior to the routine use of CT scan. Reed and colleagues in this study out of Britain acknowledge the National Institute for Clinical Effectiveness (NICE) guidelines from 2003,3 which suggest the importance of broadening the use of CT in the management of head injured children, but also note the utility of skull radiography in head injured patients when CT scanners are not available. Referencing the 2003 study by Palchak et al,4 the authors note the use of the CT scan in the United States for head injuries has increased significantly (up to 60% of patients), with only a 5-10% positivity rate. To evaluate the role of CT in head injured children, Reed et al performed a retrospective cohort analysis evaluating the effect of a change in skull x-ray policy at a United Kingdom pediatric teaching hospital. In the analysis, the emphasis on skull x-rays was reduced by restricting their use to those younger than age 1 whom they believed to be a special population. The primary outcomes included: admission rate, CT scan rate, radiation dose per head injury, and the detection of intracranial injuries. The hospital records of all head injury patients who presented during two study periods were reviewed. Each study period was one year in length; 1535 patients age 1 to 14 years were reviewed from the year prior to the change in policy and 1867 patients from one full year following the intervention. Overall, the authors found a 2:1 male to female injury pattern and an increased reliance on CT scans from 1.0% to 2.1% without evidence of an increase in rate of positive findings (25.6% vs. 25.0%) or a change in admission rate (10.9% vs. 10.1%).
Commentary
Critics of the study5 note the implications of doubling the use of CT in this young population and diligently point out that children who receive CT scans, on average, are exposed to 2mSv of radiation, while those with negative skull films and no further studies receive only 0.11mSv. Given that the small portion of those with positive skull films (and subsequent CT imaging) ended up receiving 2.11mSv, the long-term public health implications of exposing more children to higher levels of radiation without a perceived benefit needs to be evaluated further. The results of this study can be interpreted in many different ways. The authors overall recommendations suggest the abandonment of routine skull x-rays in the management of children age 1-14 years given the lack of data suggesting their utility. In light of new evidence,6 which cautions practitioners to be judicious in the use of CT scans secondary to radiation exposure, the results of this study suggest a need to re-evaluate the utility of skull x-rays as an adjunct for decision making in young children with head injuries.
Do Skull Fractures Increase Risk of Intracranial Abnormality?
Source: Andronikou S, Kilborn T, Patel M, et al. Skull fracture as a herald of intracranial abnormality in children with mild head injury: is there a role for skull radiographs? Australas Radiol 2003;47:381-385.
In this 2003 retrospective analysis of CT scans and skull radiographs done on 381 children with head injury, the authors attempt to evaluate the usefulness of skull radiographs as a clinical predictor of intracranial abnormality in children with minor head injury. The clinical records of children ranging in age from 1 day to 13 years (mean age, 6 years) from a South African Children's Hospital over a one-year period were examined. Of those included in the study, 31% of patients had intracranial abnormalities, 49% of patients were found to have a skull fracture on either CT or skull radiographs, and 49% of those had intracranial abnormalities on CT. Overall, 85% of drainable collections had an associated skull fracture, while only 54% of those with an identified intracranial abnormality had a visible fracture on plain radiograph. Unfortunately, there was not an attempt to follow those patients who received only skull radiographs and were subsequently admitted or discharged without further imaging. Therefore, the true sensitivity/specificity of skull radiographs is not clearly understood. This study, however, does suggest a limited role for skull radiographs in place of CT scans for low-risk populations or in a setting in which CT scan is not available. The authors do note that of those patients deemed to be in the "mild head injury" group with drainable collections, 100% were found to have an associated fracture.
Commentary
The degree of brain injury suffered following blunt head injury is dependent upon numerous factors. Several studies have evaluated the effect of age, mechanism of injury, and factors such as presenting GCS and other signs and symptoms on the incidence of clinically significant brain injury requiring surgical intervention. Skull fractures themselves may herald the presence of drainable intracranial hematomas, subarachnoid hemorrhages, diffuse axonal injury, and/or clinically significant skull fractures that need to be followed over time. A well described phenomenon is the "growing skull fracture of childhood" or leptomeningeal cyst. What was once a benign linear skull fracture may with time fill with cerebral spinal fluid and potentially prolapse along fracture margins, preventing fracture resolution. Those fractures that involve the dura or are associated with significant diastasis may require surgical intervention to fully heal. If such fractures are ignored, the potential exists for a growing fracture to produce large skull defects, craniocerebral erosion, and potential neurologic deficits.
Protocols that help reduce the pretest probability in intracranial abnormalities may create a role for the use of skull radiographs. If done correctly, existing data suggest that patients with low-risk injuries could be appropriately risk stratified into CT or no CT groups with the aid of skull radiographs; this would reduce the need for higher levels of radiation exposure.
How Much Risk is Associated with Head CTs?
Source: Brenner DJ, Elliston CD, Hall EJ, et al. Estimated risk of radiation induced fatal cancer from pediatric CT. AJR Am J Roentgenol 2001;176:289-296.
In this study, Brenner and colleagues use predetermined estimations of body organ exposure to radiation from various types of computed tomography based on age at exposure to predict the lifetime risk of cancer mortality. The authors' estimation of the lifetime cancer mortality risk in pediatric populations as a function per unit dose for a given age, clearly illustrates the impact different types of CT scans can have on the lifetime risk of death from cancer from a given organ system. Brenner and co-workers suggest that it is the combination of higher than average doses of radiation and lifetime risks related to those doses in pediatric CT that is thought to result in a sharp increase in the estimated risk relative to adult CT. Using a number of assumptions related to the predicted incidence of head CTs in the pediatric population, Brenner et al predict the total number of deaths attributable to 1 year of CT examinations in the United States to be ~700 from head examinations and ~1800 from abdominal CTs. Using these assumptions, it is estimated that the lifetime cancer mortality risk attributable to the radiation exposure from a single head CT in a 1-year-old child is approximately 1 in 1500. With more than 600,000 abdominal and head CT scans performed per year in children younger than age 15, Brenner predicts 500 attributable deaths.
Commentary
Diagnostic radiography was revolutionized with the advent of CT in the 1970s. As the indications for its use have broadened, the number of scans per year has climbed markedly. Current estimates suggest that more than 60 million scans per year are obtained, and that at least 4 million of these are in children. Unfortunately, as the clinical applications for this study increase so does the lifetime risk of cancer. This is of particular concern in younger populations, as the predicted lifetime risk of radiation exposure can only increase with time.6 This evaluation makes a good argument for limiting the indications of CT. However, what also is clear is that CT has revolutionized the diagnostic and therapeutic capabilities of medicine. Clearly, what was once thought of as an indispensable diagnostic modality now needs to be critically evaluated for its clinical implications on a case-by-case basis.
What Guidelines Aid Management of Mild Head Injury in Children Younger than Age 2?
Source: Schutzman SA, Barnes P, Duhaime AC, et al. Evaluation and management of children younger than two years old with apparently minor head trauma: proposed guidelines. Pediatrics 2001;107:983-993.
Schutzman and colleagues present the most concise evaluation of head injury in infants < age 2 available in the literature. Born out of expert consensus on available evidenced-based reports, Schutzman et al developed a management strategy that categorizes children younger than age 2 into 4 subgroups based on the estimated risk of intracranial injury (ICI). This analysis also provides answers to 10 frequently asked questions relevant to head injuries in this population and describes a clear-cut strategy for the management of patients based on their predicted severity of injury. More than 404 articles were reviewed by an expert panel composed of 9 full-time academic faculty members with demonstrated expertise in pediatric head injury. The panel included 4 pediatric emergency medicine physicians, 1 emergency physician, 2 pediatric neurosurgeons, 1 pediatric neuroradiologist, and 1 general pediatrician. Each panel member reviewed the selected literature; consensus was reached in a closed meeting on the specific clinical questions identified, and a management algorithm was developed.
Identified questions of interest focused on the issues related to imaging and disposition. In cases in which data were insufficient to definitively answer proposed questions, expert consensus was used. The following section provides abbreviated answers to the questions addressed by the panel, as described by Schutzman et al.
1. What are the indications for CT?
CT is the standard for the diagnosis of acute ICI. The authors note ICI is present in this population of minor head injuries ~3-6% of the time. The clinical predictors of ICI noted in the literature include, but are not limited to, the presence of a skull fracture (SF), focal neurologic findings, scalp swelling, younger age, inflicted injury, and head injury without a clear history of mechanism. The authors make special note of the increased incidence of injury in patients <3 to 6 months of age. Loss of consciousness (LOC) and vomiting have not been shown to be strong predictors of ICI, but consensus advocates for concern in these patients. The presence of SFs alone are thought to be one of the strongest predictors of ICI, present in 60-100% of ICIs.
2. What are the indications for skull radiographs?
Skull radiographs (SR) are capable of identifying SFs. The younger the patient, the higher the risk of SF. According to the literature, SFs are present in 6-30% of outpatients evaluated for head injury and in 80-100% of patients with scalp hematomas.
3. If a fracture is noted on SR, should a CT be obtained?
Here, the authors point out the strong association of SFs with ICI; 15-30% of SFs have an associated ICI. Since the presence of a SFs is a strong indicator of ICI, CT is indicated for a child with a SF.
4. If the CT is read as normal, which children may be discharged from the hospital?
In three studies, with a total of 261 patients, the incidence of subsequent deterioration in children with a normal CT was zero.
5. If a linear SF is diagnosed but no ICI is present, what are the criteria for discharge?
Six studies with a total of 349 patients indicate there was no incidence of clinical deterioration in patients with isolated linear SF.
Management Guidelines. The second portion of this consensus statement addresses a management strategy for this young population. Here, the authors define 4 basic clinical subgroups based on estimated risk and suggest diagnostic strategies. The 4 basic subgroups are described in the following section (see Figure 1).
![]() |
High-Risk Group. This group has signs and symptoms concerning for ICI, and CT is indicated. High-risk patients include those with depressed mental status, focal neurologic findings, signs of depressed or basilar SF, SF noted on exam or SR, irritability, or bulging fontanel. Although not supported by the literature, consensus supports the use of CT in those patients with documented seizure activity, LOC > 1 minute, and persistent vomiting. Children <3 months of age are considered to have a relatively high risk of ICI, with minimal evidence on clinical assessment; therefore, a low threshold for CT should exist.
Intermediate Risk Group. This group is further subdivided into two groups. In the first group, either CT or observation are valid options, while an allowance for imaging with CT and/or SR or observation is suggested for the second subgroup.
The first subgroup in this cohort is those children with clinical indicators of possible brain injury; they are defined here as children with: 3-4 episodes of vomiting, transient LOC <1 min, history of lethargy or irritability behavior not at baseline, and nonacute SF (>24hours old). The authors suggest that when >1 of the criteria are present, LOC is >15-30 seconds, behavior change > 30 minutes, or the child is younger, CT should be strongly considered. In those patients who do not have a CT performed, observation for 4-6 hours is recommended. Any progression of symptoms should trigger a CT.
The second subgroup includes those patients with concerning or unknown mechanism of injury who have findings on physical exam that may indicate an underlying SF. Included here are those children with a higher force mechanism (high-speed motor vehicle accident, ejection, falls >3-4 feet), falls onto hard surfaces, large boggy scalp hematomas (particularly in the temporoparietal region), unwitnessed trauma with a possibility of significant mechanism, and those with a vague or absent history of injury with indicative signs and symptoms. Here the recommendation is to obtain imaging either CT or SR, based on the given clinical scenario. A multitude of factors may need to be taken into consideration, including the availability of the diagnostic modality, expertise of imaging interpretation, and need for sedation. Those patients not imaged should be observed for 4-6 hours.
Low-Risk Group. This group includes those with trivial injuries who are thought to have a very low likelihood of ICI. Children in this category have experienced a very low energy mechanism of injury (e.g., fall <3 feet) and have no signs or symptoms 2 hours post injury. In this population, age >3-6 months is indicative of a reassuring prognosis.
Commentary
Here Schutzman and colleagues were able to use available evidence in conjunction with expert consensus to successfully answer questions pertinent to the management of head injuries in children younger than age 2. The end result is a set of clearly defined elements of risk with suggested management strategies. Children younger than age 2 present with a unique set of elements that make for a difficult diagnostic challenge. This population is often more difficult to assess clinically, and asymptomatic occult ICI is frequently present. The youngest patients who are at greater risk are oftentimes harder to image without sedation; furthermore, this age group in particular is at greater risk for non-accidental trauma, which has its own set of social barriers. To date, this is the most concise set of recommendations concerning the management of these patients. Lacking in this analysis is consideration of the new appreciation of risk from CT. Due largely to the lack of evidence at the time of publication, no mention is given to the implication of high levels of radiation exposure inherent in the proposed guidelines. Although this consensus statement reflects only the expert interpretation of the evidence of the time, it stills serves as the best extrapolation to date.
Scalp Hematomas As Important Predictors of Skull Fracture and Associated Injury
Source: Greenes DS, Schutzman SA. Clinical significance of scalp abnormalities in asymptomatic head-injured infants. Pediatr Emerg Care 2001;17:88-92.
Greenes and Schutzman, now recognized as leaders in the realm of pediatric head injury, conducted an eloquent prospective cohort study of asymptomatic head injured children ages 0-24 months. The intent was to identify clinical indicators of high-risk skull fractures and associated ICI. The study cohort included 433 asymptomatic patients presenting to an academic pediatric emergency department following a preceding head injury. Patients were considered to be asymptomatic if they lacked clinical evidence of head injury (including loss of consciousness, lethargy, irritability, seizures, 3 or more episodes of vomiting, abnormal vital signs suggestive of increased intracranial pressure, or focal neurologic findings), depressed skull fracture, or a basilar skull fracture. This presumably low-risk population was evaluated for the presence, size (none, small, moderate, and large), and location (frontal, parietal, temporal, and occipital) of scalp hematomas. Treating physicians followed recommended diagnostic guidelines, which encourage CT scans as the initial study for all symptomatic patients and skull radiographs as the initial study for asymptomatic patients. Any patient with a documented SF on initial skull radiographs was to have a CT scan.
Of the 422 patients evaluated, 45 (11%) were found to have a SF and 13 (3%) an ICI, all of which had an associated SF. Noting the clear relationship between location and size of hematoma with the presence of skull fracture, the authors used the collected data to develop a decision rule to aid in the diagnosis of SF and ICI. With a sensitivity of 0.98 and a specificity of 0.49 for the presence of SF, their decision rule correctly identified 100% of the ICIs. The clinical decision rule is listed in Table 1.
![]() |
By adding the points associated with each column, patients' risk of skull fracture and/or ICI was predicted based on scores ranging from 0-8. The prevalence of both SF and ICI increased with increasing scores. No evidence of ICI was found with a score of 3 or less. One skull fracture was noted with a score of 1 (4%). The incidence of SF increased with rising scores, with 100% of patients with a score of 7 or 8 demonstrating evidence of SFs. Similarly, a score > 3 predicted a nearly linear relationship with ICI.
Commentary
Greenes and Schutzman did an excellent job elucidating the relationship between the size, appearance, and location of scalp hematomas with patient age and the likelihood of SF and ICI. Of note, there was a strong association with large boggy hematomas and those located in the parietal and temporal regions. There does not appear to be an increased risk of injury with frontal hematomas. The authors note special concern in patients younger than age 1, with emphasis on definitive imaging in infants younger than 3 months without a clearly described trivial mechanism. This tool is of great utility in the evaluation of the asymptomatic, well appearing child younger than age 2. Unfortunately, the results of this study are hampered by the limited number of ICIs and SFs. However, it is now clear that a relationship exists between the size and location of hematomas. As suggested by the authors, applying this rule clinically, children with scores of <3 may not need any imaging, scores ranging from 3-4 should initially have skull radiographs, and those with scores >4 should go directly to CT. The result is a useful clinical prediction tool that allows for a fraction of the radiation exposure in a subset of patients with little to no risk of ICI.
In Children Older than Age 2, who needs a CT Scan?
Source: Halley MK, Silva PD, Foley J, et al. Loss of consciousness: when to perform computed tomography. Pediatr Crit Care Med 2004;5:230-233.
Halley and colleagues attempt to define the importance of the Glasgow Coma Score (GCS), physical examination, and history of loss of consciousness (LOC) or amnesia following blunt head injury as clinical predictors for intracranial injury requiring neurosurgical intervention in children age 2-16. In a prospective analysis of 98 children age 2-16 years with a GCS of 13-15 and LOC or amnesia, a standardized physical examination including a brief neurologic evaluation and head CT were used to assess the ability of clinical exam to predict the need for neurosurgical intervention. The initial operating hypothesis was that GCS together with history and physical exam was adequate to predict clinically important findings on CT.
Of the 98 subjects enrolled, 13 (13.3%) had evidence of intracranial injury. Four of these subjects had a negative physical exam (GCS 15, no scalp findings, no hemotympanum, normal neurologic exam). Although admitted and monitored with subsequent CT scans, none of the 4 subjects required neurosurgical intervention. In total, only two of the 13 subjects required neurosurgical intervention. Overall, a positive physical exam (any abnormal finding on the standardized examination or GCS <15) had a sensitivity of 69% and a specificity of 40% with a positive predictive value of 15% and negative predictive value of 89%. An abnormal physical exam increased the probability of an abnormal CT from 0.13 pretest to 0.15 posttest, while a normal physical exam increased the probability of a normal CT from 0.87 pretest to 0.90 posttest.
Commentary
This study attempted to highlight the clinical utility of a focused physical exam, a patient's GCS, and historical LOC and/or amnesia in the evaluation of head injured children. Although the authors intent was to define clinical criteria that could be used to limit the use of CT, their limited criteria would have resulted in ~4% missed injury rate. None of the "missed" subjects ultimately would have required subsequent neurosurgical intervention; however, as the authors diligently point out, the long-term effects of intracranial injuries with regard to cognition and developmental delays is not well studied. Interestingly, each of these patients demonstrated complaints of headache and variable emesis. This study was limited by a number of obvious factors. By limiting their historical values to LOC and/or amnesia, the authors eliminate the potential clinical importance of persistent emesis, increasing headache, and other such signs and symptoms. Such historical findings have been shown in adult patients to be potential clinical indicators of significant ICI. As a result, this study suggests the need for CT in all patients age 2-16 with LOC, amnesia, and/or GCS <15, regardless of physical and historical findings outside of these limited criteria.
Summary
With mounting evidence that suggests the potential harm of the liberal use of CT, clinical studies that evaluate the combination of both skull radiography and head CTs in the evaluation of head injury need to be investigated. Existing data clearly support the use of CT in any child with clinical evidence of severe head injury. Although various definitions exist, any child with a prolonged LOC, evidence of basilar or depressed skull fracture, large scalp hematomas in either the parietal or temporal regions, persistent emesis, GCS <15, seizure activity, focal neurological deficits, or any child <3-6 months of age without a known trivial mechanism of injury is considered to be at high risk and should be considered strongly for definitive CT.
Skull radiography still may have limited utility in those patients with suspected trivial mechanisms who lack clinical evidence of injury. Once identified, skull fractures should prompt the use of CT as well as subsequent outpatient evaluations for evidence of fracture evolution. Once thought to be a harmless diagnostic modality, CT should now be used with deliberate caution for the evaluation of trivial head injuries. Patients and their caregivers should be counseled on the associated risks and benefits of the available diagnostic options for their given injuries.
References
1. Haydel MJ, et al. Indications for computed tomography in patients with minor head injury. N Engl J Med 2000;343:100-105.
2. Stiell IG, et al. The Canadian CT Head Rule for patients with minor head injury. Lancet 2001;357:1391-1396.
3. Leaman AM. The NICE guidelines for the management of head injury: the view from a district hospital. Emerg Med J 2004;21:400.
4. Palchak MJ, et al. A decision rule for identifying children at low risk for brain injuries after blunt head trauma. Ann Emerg Med 2003;42:492-506.
5. Tasker RC. Skull x rays, CT scans, and making a decision in head injury. Arch Dis Child 2005;90:774-775.
6. Brenner DJ, Hall EJ. Computed tomography an increasing source of radiation exposure. N Engl J Med 2007;357:2277-2284.
Pediatric head injury is the leading cause of trauma mortality in children.Subscribe Now for Access
You have reached your article limit for the month. We hope you found our articles both enjoyable and insightful. For information on new subscriptions, product trials, alternative billing arrangements or group and site discounts please call 800-688-2421. We look forward to having you as a long-term member of the Relias Media community.