Introduction: Children with isolated skull fractures (SF) diagnosed with a computed tomography (CT) scan do not require hospitaliza- tion if they are neurologically normal. The aim of this study was to assess whether CT scans are necessary to confirm a suspected SF or whether children appearing well can be managed as outpatients without radiographic examinations.
Methods: This retrospective study analyzed data from children up to 16 years with a radiographically confirmed SF in 2014 and 2015. The variables analyzed included demographic data, the child’s general condition, the Glasgow Coma Scale (GCS), the management of radiographic examinations and observations, and the necessity of a neurosurgical intervention.
Results: A total of 68 patients were analyzed; their mean age was 3 years (range 4 days – 15 years), and most fractures were diagnosed with skull radiographs (60.3%; 41/68), followed by CT scans (36.7%; 25/68). Fifty-five children (80.9%) appeared well with a GCS of 15, although in two infants, intracranial hematomas were found with no need for intervention. In contrast to children with GCSs of 13 and 14, four of six had intracranial injuries, two of whom required neurosurgical interventions.
Discussion: When a SF is suspected, children older than one year who appear well with a GCS of 15 and a normal neuro- logical exam result might be managed safely as outpatients if parents are reliable and understand return-for-care criteria. In these cases, a radiographic confirmation of the clinically suspected SF does not seem necessary for acute management.
Head traumas in children are common injuries ac- counting for up to 500,000 emergency department (ED) visits annually in the United States . How to manage these injuries thus has major implications. Consensus ex- ists on necessary computed tomography (CT) examinations and inpatient observations for the management of pediat- ric head traumas that are moderate (Glasgow Coma Scale (GCS) between 9–12) and severe (GCS between 3–8), yet there is controversy regarding the management in mild (GCS between 13–15) pediatric head trauma (MPHT) [1-4].
An isolated skull fracture (SF) without intracranial injury is identified in approximately 10% of all traumatic neuroimaging, which often results in hospitalization be- cause such children are usually grouped into high-risk MPHT due to the potential risk of clinical decompensa- tion and evolving intracranial hemorrhage [3-9]. Children with a suspected SF regularly undergo CT scans to rule out intracranial hemorrhage or injury; however, the risk of finding such intracranial abnormality is 10-30%. Because neurosurgical interventions are rare, the necessity of hos- pitalization has increasingly been brought into question [1,3,5,7,10,11]. Several recent studies suggest outpatient management in neurologically normal children with an iso- lated SF in the CT scan [1,3,5,7,11].
In our institution, CT scans are performed restric- tively because of the risk of radiation-induced malignancy for patients; CT scans are indicated for children younger than three months with a head trauma and a clinical high suspicion of a skull fracture, children with altered neuro- logical status, or children with a coagulopathy disorder . In children older than three months who present in a good general condition with GCS 15 and clinical signs of a skull fracture (skull hematoma, in particular non-frontal large and boggy hematoma), the skull is radiographed and if the radiographs are consistent with a skull fracture, the child is admitted for inpatient neurological observation for 72 hours, during which CT scans are only performed if the child deteriorates neurologically .
The aim of this study was to assess whether a sus- pected skull fractured can be safely managed in outpatient care and without radiographic examinations in children who appear well with a GCS of 15 and a normal neurological exam result.
Material and methods Study design
This study was designed as a retrospective, sin- gle-center analysis of children with a confirmed skull frac- ture between 2014 and 2015. The study was approved by the local ethics committee.
Our study consists of data from children up to 16 years with a skull fracture diagnosed in our ED. Only ra- diographic confirmed fractures were eligible for inclusion in the study, all of which were assessed by pediatric radiol- ogists. Skull radiographs included both lateral views and anterior-posterior or Towne projection. Another inclusion criterion was a signed parental general consent at the initial ED presentation that allowed data to be used for research thereafter. Children were excluded if they were directly transferred to our intensive care unit, had multiple traumas with involvement of other organ systems, or were referred to our institution due to suspected non-accidental trauma.
We first determined the variables to be analyzed. These included demographic data, the child’s general con- dition, the GCS, the management of radiographic examina- tions and observations, the necessity of a neurosurgical intervention and a child’s return after discharge due to complications. A child who appeared well was defined as a child assessed to be in good general condition with a GCS of 15, who was consolable and had normal skin color, spon- taneous motor activity, and interaction.The data was ano- nymized and transcribed into SPSS. Here, the data are de- scribed as frequencies or means with range as appropriate.
A total of 68 patients were eligible and analyzed; they had a mean age of 3 years (range 4 days – 15 years) and most fractures were diagnosed with skull radiographs (60.3%; 41/68), followed by CT scans (36.7%; 25/68), and magnetic resonance imaging (MRI) or ultrasound (1.5%; 1 case each).
Overall, the most frequent trauma mechanism was falls (88.2%; 60/68), followed by the head struck by an object (7.4%; 5/68) and traffic accidents (4.4%; 3/68) [Table 1].
Table 1: Demographics, diagnostics and outcomes in skull fractures in children who appear well and children who do not appear well.
A MPHT was diagnosed in almost all children (94%); 59 had a GCS of 15; two children presented a GCS of 14, and three others a GCS of 13. Two further children showed a GCS of 11 and two others a GCS of 7. The children with a GCS of 15 appeared well in almost all cases (93.2% 55/59), while the remaining four children showed either an epidural (3) or a subdural hematoma (1). All 55 children who appeared well had an uneventful observation and no unplanned return visit after discharge. Of the six children with a GCS of 13 to 14, four had additional injuries to the SF, including three with an epidural hematoma and one with a cerebral contusion.
Observation in the intensive care unit was neces- sary for seven children, of whom two required neurosurgi- cal interventions. The most frequent injury mechanisms in these cases were falls: Two infants fell together with their mothers (a 2-month and a 3-month-old child); one boy (1 year old) fell from his father’s shoulders (initial GCS 13, posttraumatic focal seizure) and required a neurosurgi- cal intervention; a four-year-old boy fell from his loft bed, and a seven-year-old girl fell from a height of two meters while playing (GCS 7, neurosurgical intervention). Traffic accidents accounted for two injuries: A nine-year-old boy was hit by a car while riding his bike and a 15-year-old girl was hit by a car while crossing the street. All seven children arrived in a reduced general condition with a GCS from 7 to 14; six of the seven had several episodes of vomiting and the one without vomiting was inconsolable (two-month-old infant).
Children, who appear well, are older than one year, and have a GCS of 15 as well as a normal neurological exam result in the setting of a suspected SF might be managed safely as outpatients if parents are reliable and understand return-for-care criteria. In these cases, a radiographic con- firmation of the clinically suspected SF seems unnecessary for acute management.
Half of the children in this study were less than one year old. In the age group up to three months, the typical injury mechanism was falling together with their caretaker, and in the age group up to one year, falling because of the in- fant’s ability to turn. Half of the children in the age group up to three months who were included in this study had either an epidural or subdural hematoma or a cerebral contusion. These infants are difficult to assess clinically and might only show subtle signs and symptoms, so when a skull fracture is suspected, current guidelines suggest a conservative approach including CT scans . For children six months and older, moreover, there seems to be a very low risk of complications if they appear well and only present with a buggy hematoma with no sign of neurological anomalies or non-accidental trauma and no clinical signs of fractures of the base of the skull (raccoon eye, hemotympanum).
The symptoms that predict intracranial injuries are seizure, altered mental status, focal neurolog- ical deficit, and signs of basilar skull fracture . Another predictor for an abnormal CT result is vomiting, yet its val- ue is less clear [2,9]. Funderò et al. found vomiting to be a significant risk factor for an abnormal CT result, whereas Schacher et al. considered vomiting an insignificant factor for intracranial injuries [13,14]. We found an overall rate of 13.2% of vomiting, 14.5% in the subgroup of children who appeared well and 7.7% in the other subgroup. Thus, we have concluded that the child’s general condition seems to determine whether the combination of a SF with vomiting is an indicator of intracranial injuries.
Our data showed that a substantial majority (80%) of the children with SF appeared well, had a GCS of 15, and had an uneventful outcome. In this subgroup, however, a two-month-old infant who had fallen down stairs present- ed a 4-mm-wide subdural hematoma confirmed by CT and one infant had an epidural hematoma diagnosed by MRI in the course of investigating a possible non-accidental trau- ma. These two infants with anomalies in the CT scan were hospitalized for three days prior to discharge and no neuro- surgical or clinical interventions were necessary. Five fur- ther children with a GCS of 13-14, also regarded as MPHT, did not appear well and therefore underwent a CT scan. In three children, an epidural hematoma was evident and, in another child, a cerebral contusion. Three of these children required observation in the intensive care unit. Of the 26 children who had a CT scan or MRI, slightly more than one third (38.5%) were diagnosed with intracranial injuries, which is in line with previous research. It is likely, howev- er, that some intracranial injuries were missed in the group of children who appeared well because these examinations were only performed in 38% of all children with SFs.
Outpatient management for children with SF with no intracranial injuries is recommended for children older than one year [1,3,5,7,10,11]. Thus, if there is no difference in acute management between children with and without a skull fracture, it is questionable whether radiograph- ic examinations are justified. A question then arises as to whether a clinical examination would also be sufficient in children who appear well and have suspected skull frac-tures. In our subgroup of children who appeared well with SFs, a substantial majority (41, 74.5%) underwent only a plain radiograph and no CT scan, which means that we did not have the potential additional findings provided by a CT scan; however, this did not impact acute management. Here it should also be noted that even if intracranial injuries are found in a CT, the chance of a complication requiring emer- gency neurosurgery or clinical intervention is extremely low . Nevertheless, one complication feared after a head trauma is the development of an epidural hematoma and its potential risk of death. Such a complication usually occurs immediately following trauma, yet there are numerous re- ports of delayed and progressive hematomas three or more days following the trauma [15,16]. This means that even if a child is hospitalized for 24 hours, the child’s condition may deteriorate later, which has led us to conclude that hospital admission should be reserved for symptomatic children or those suspected of non-accidental trauma. These points to a need to pursue an optimized indication of radiographic examinations, especially CT scans and plane radiographs, because such an optimization would reduce radiation ex- posure as well as lower general medical costs and consulta- tion time.
The major drawbacks of this study are its retro- spective design and the small number of complicated skull fractures, which limit the interpretation and generalization of our results. In addition, some skull fractures were proba- bly missed due to our restrictive approach. However, it may reasonably be assumed that most of these children experi- enced no complication because none of them returned or were secondarily referred to our center for further treat- ment. Finally, we focused only on the short-term, not long- term, outcomes of the children.
When a SF is suspected, children older than one year who appear well with a GCS of 15 and a normal neurological exam result might be managed safely as outpatients if par- ents are reliable and understand return-for-care criteria. In these cases, a radiographic confirmation of the clinically suspected SF does not seem necessary for acute manage- ment.
The author has no competing interests to declare
No sources of funding were used.
I thank Sandra Anusic for her assistance in data collection.