See the corresponding editorial, DOI: 10.3171/2014.3.PEDS14104.
DOI: 10.3171/2014.5.PEDS13414 ©AANS, 2014
Pediatric skull fractures: the need for surgical intervention, characteristics, complications, and outcomes Clinical article Christopher M. Bonfield, M.D.,1 Sanjay Naran, M.D., 2 Oluwaseun A. Adetayo, M.D., 2 Ian F. Pollack, M.D.,1 and Joseph E. Losee, M.D. 2 Departments of 1Neurological Surgery and 2Plastic Surgery, University of Pittsburgh, Pennsylvania Object. Head trauma is a common cause of morbidity and mortality in the pediatric population and often results in a skull fracture. Pediatric skull fractures are distinct from adult fractures. Pediatric fractures have a greater capacity to remodel, but the pediatric brain and craniofacial skeleton are still developing. Although pediatric head trauma has been extensively studied, there is sparse literature regarding skull fractures. The authors’ aim was to investigate the characteristics, injuries, complications, and outcomes of the patients in whom surgical intervention was needed for skull fractures. Methods. The authors performed a retrospective review of patients presenting to the emergency department of a pediatric Level I trauma center between 2000 and 2005 with skull fractures. Patient demographics, mechanism of injury, associated injuries, fracture bone involvement, surgical intervention, complications, and outcomes were analyzed. Groups treated nonoperatively, for skull fracture repair, and for traumatic brain injury were compared. Results. A total of 897 patients with a skull fracture were analyzed. Most patients (n = 772, 86.1%) were treated nonoperatively (Non-Op group). Fifty-eight patients (6.5%) underwent repair of the fracture (Repair group) and 67 (7.5%) required intervention for treatment of traumatic brain injury (TBI group). The Non-Op group was significantly younger, and the TBI group had a lower initial Glasgow Coma Scale (GCS) score. A fall (51.2%) was the most common mechanism of injury in the Non-Op group, whereas a motor vehicle crash (23.9%) and being hit in the head with an object (48.2%) were most prevalent in the TBI and Repair groups, respectively. Associated injuries were seen in all 3 groups, with brain injury (hematoma) being the most common. Frontal bone fracture was seen most in the Repair and TBI groups, and the parietal bone was the most frequent bone fractured in the Non-Op group. Patients in the TBI group were much more likely to have 2 or 3 skull bones fractured. In the Repair group, 36.2% had a complication (38.0% intervention related and 62.0% trauma related), but no patient had a worsening of their neurological status. In the TBI group, 48.7% of the patients suffered a complication, the vast majority (90.6%) of which were related to the trauma. Conclusions. The majority of pediatric skull fractures can be managed conservatively. Of those requiring surgical intervention, fewer than half of the surgeries are performed solely for skull fracture repair only. Patients hit in the head with an object or involved in a motor vehicle crash are more likely to need surgical intervention either to repair the skull fracture or for TBI management, respectively. Frontal bone fractures are more likely to necessitate repair, and those patients treated for TBI have a greater incidence of 2 or 3 bones involved in the fracture. Complications occurred but most were related to underlying trauma, not the surgery. No patients who underwent intervention for repair of their skull fracture only had a worsening of their neurological status. (http://thejns.org/doi/abs/10.3171/2014.5.PEDS13414)
Key Words • pediatric • skull fracture • head injury • trauma • surgical intervention
H
ead trauma is common in the pediatric population and is an important cause of morbidity and mortality in the United States, with an estimated incidence
Abbreviations used in this paper: EVD = external ventricular drain; GCS = Glasgow Coma Scale; ICU = intensive care unit; LOS = length of stay; MVC = motor vehicle crash; TBI = traumatic brain injury.
J Neurosurg: Pediatrics / June 6, 2014
of 250 per 100,000 per year. It accounts for over 7000 deaths, 60,000 hospitalizations, and 600,000 emergency department visits annually among American children.8,12 In children, between 10%–30% of head injuries result in skull fracture,7,9 many with associated brain injury.2,13 Surgical intervention is largely performed in cases of skull fracture depression, frontal sinus involvement, and underlying mass lesion. In certain instances, especially in infants, conserva1
C. M. Bonfield et al. tive management has been advocated as treatment.5,15 Pediatric skull fractures remain distinct from their adult counterparts. Pediatric skull fractures have a greater capacity to heal and remodel, but the pediatric brain and craniofacial skeleton are also still developing, which puts the children at risk for unique complications, such as growing skull fractures. Although pediatric head trauma has been extensively studied, there is sparse literature regarding skull fractures, the rate of operative management, and outcomes following surgical intervention in this population. Our goal was to investigate the characteristics, injuries, indications for intervention, and outcomes of the pediatric patients in whom surgical intervention was needed for cranial vault skull fractures.
Methods
The study was approved by the Institutional Review Board at the University of Pittsburgh. A database of all patients diagnosed with a skull fracture at the Children’s Hospital of Pittsburgh from 2000 to 2005 was searched. All patients diagnosed with a skull fracture were included regardless of treating specialty, treatment modality, or need for hospital admission. Patient demographics, mechanism of injury, associated injuries, fracture bone involvement, surgical intervention, complications, and outcomes were recorded. Comparison was made between individuals who were treated nonoperatively (Non-Op group), those surgically treated for skull fracture repair only (Repair group), and those surgically treated for traumatic brain injury (TBI) with or without repair of a skull fracture (TBI group). Indications for intervention in the Repair group included displaced or depressed fracture, open fracture, frontal sinus fracture, and cosmetic deformity. Patients in the TBI group had intervention for treatment of the underlying trauma, and the group included such operations as hematoma evacuation, decompressive hemicraniectomy, external ventricular drain (EVD) placement. Skull fractures may have been repaired in the TBI group, but that was not the principal indication for surgery. In general, for children younger than 2 years of age, sutures were used to affix the bone. For children older than 2 years of age, titanium plates and screws were used. Data analysis was performed using SPSS version 17 (SPSS, Inc.). Chi-square tests were used for the betweengroup comparisons of categorical variables. Since age, length of hospital stay, and Glasgow Coma Scale (GCS) score were not normally distributed, Kruskal-Wallis tests were used for the comparisons of continuous variables. Mean values presented ± SD.
Results Patient Characteristics
A total of 897 patients presented to Children’s Hospital of Pittsburgh between 2000 and 2005 with a diagnosis of a skull fracture. Of these patients, 772 (86.1%) were treated nonoperatively (Non-Op group). The remaining patients were grouped according to the indication for their surgical intervention. Fifty-eight patients (6.5%) un-
2
derwent repair of the fracture (Repair group) for fracture elevation, frontal sinus repair, open fracture debridement, or cosmetic repair. Sixty-seven patients (7.5%) required intervention for treatment of a TBI (TBI group) including hematoma evacuation, EVD placement, or decompressive craniectomy. Patient demographics are listed in Table 1. In all 3 groups, the majority of patients were male (63.5% in the Non-Op, 69.0% in the Repair, and 65.7% in the TBI group) and Caucasian (Non-Op 82.3%, Repair 81.0%, and TBI 85.1%). The mean age at the time of evaluation was significantly younger in the Non-Op group (5.5 ± 5.0 years, p < 0.001) compared with the Repair (8.5 ± 4.8 years) and TBI (8.6 ± 5.2) groups. Hospital Stay Data
Hospital stay data are also reported in Table 2. The initial GCS score was significantly lower in the TBI group (6.80 ± 4.5, p < 0.001) in relation to the Non-Op (13.6 ± 3.4) and Repair (12.4 ± 4.6) groups. In the TBI group, 63 patients (94.0%) were admitted to the intensive care unit (ICU) and 42 patients (62.7%) were intubated. In the Repair group, 24 patients (41.4%) were admitted to the ICU and 12 patients (20.7%) were intubated. The Non-Op group had the lowest rates of ICU admissions (19.4%) and intubations (10.0%). The length of stay (LOS) also varied between the groups. The TBI group had the longest LOS (16.0 ± 13.0 days, p < 0.001), followed by the Repair and Non-Op groups (5.3 ± 7.0 days and 2.1 ± 3.6 days, respectively).
Mechanism of Injury
The mechanisms of injury varied among the groups (Table 3). The most common injuries sustained by the Non-Op group were fall (n = 395, 51.2%), object to head (n = 130, 16.8%), and motor vehicle crash (MVC; n = 85, 11.0%). In the Repair group, the most common mechanisms of injury were object to head (n = 28, 48.2%), fall (n = 8, 13.8%), and MVC (n = 6, 10.3%). The most common injuries in the TBI group were MVC (n = 16, 23.9%), fall (n = 12, 17.9%), object to head (n = 12, 17.9%), and pedestrian hit by motor vehicle (n = 8, 11.9%). The mechanisms resulting in the highest percentage of surgical intervention (Table 4) were bicycle accident (33.3% of total bicycle accidents), object to head (23.5% of total objects to head), and pedestrian hit by motor vehicle (22.0% of total pedestrians hit by motor vehicle).
Fracture Location
Location of the fractures was also recorded (Table 5). Overall, the most common location was the parietal bone (n = 269, 30.0%), followed by the frontal bone (n = 242, 27.0%). However, the most common bone fractured differed among groups (p < 0.001). The parietal bone was the most common fracture location in the Non-Op group (n = 251, 32.5%), whereas the frontal bone was the most injured in the Repair and TBI groups (n = 36, 62.1%, and n = 18, 26.9%, respectively). In the repair group, 8 (22.2%) of 36 required surgery for frontal sinus repair and 28 (77.8%) of 36 had treatment for elevation of a deJ Neurosurg: Pediatrics / June 6, 2014
Surgery for pediatric skull fractures TABLE 1: Summary of patient demographics Treatment Group (%) Variable sex* male female total race† white black other/unknown age (yrs)‡ mean range SD
Non-Op
Repair
TBI
All
490 (63.4) 282 (36.5) 772 (100)
40 (69.0) 18 (31.0) 58 (100)
44 (65.7) 23 (34.3) 67 (100)
574 (64.0) 323 (36.0) 897 (100)
635 (82.3) 86 (11.1) 51 (6.7)
47 (81.0) 8 (13.8) 3 (5.2)
57 (85.1) 6 (9.0) 4 (6.0)
739 (82.4) 100 (11.1) 58 (6.5)
5.5 0.1–21.7 5.0
8.5 0.1–17.7 4.8
8.6 0.1–17.6 5.2
5.9 0.1–21.7 5.1
* p = 0.67. † p = 0.92. ‡ p < 0.001.
pressed fracture and cosmetic indications. The 18 patients in the TBI group had underlying brain injuries that necessitated intervention. The TBI group had the highest rate of injuries involving 2 (23.9%) and 3 (7.5%) bones. The Non-Op group (2 bones in 11.0% and 3 bones in 1.4%) and the Repair group (2 bones in 13.8% and 3 in bones 0.0%) had lower rates of multiple bone involvement in fractures. The fractured bones most likely to result in a patient undergoing surgical treatment (Table 6) were 3 bones involved (31.3% of total 3 bone–involved injury), frontal bone (22.3% of total frontal bone injury) and 2 bones involved (22% of total 2 bone–involved injury). Associated Injuries
Many patients sustained injuries to other areas of body besides the skull. Associated injury rates are reported in Table 7. Intracranial hemorrhage (epidural, subdural, subarachnoid, or intraparenchymal) was the
most common associated injury in all 3 groups (Non-Op group: n = 326, 42.2%; Repair group: n = 23, 39.7%; and TBI group: n = 57, 85.1%). The next most encountered associated injury was orthopedic (n = 82, 10.6%) in the Non-Op group, face (n = 9, 15.5%) in the Repair group, and cardiac/pulmonary (n = 19, 28.3%) in the TBI group. Procedures
The types of intervention in the surgical groups were reviewed. All fractures, except one, in the Repair group were depressed (n = 57, 98.3%), and 21 fractures (36.2%) were open. These fractures underwent procedures for elevation, repair, or debridement. The remaining patient underwent the repair of a growing skull fracture 8 months after the initial injury. In the TBI group, the most common procedure was the insertion of an EVD (77.6%). The most common open surgical intervention was a decompressive craniectomy (22.4%), followed by epidural
TABLE 2: Hospital stay data Treatment Group (%) Variable GCS score on admission* mean range SD ICU admission* intubated* LOS (days)* mean range SD
Non-Op
Repair
TBI
All
13.6 3–15 3.4 150 (19.4) 77 (10.0)
12.4 3–15 4.6 24 (41.4) 12 (20.7)
6.8 3–15 4.5 63 (94.0) 42 (62.7)
13.0 3–15 4.0 237 (26.4) 131 (14.6)
2.1 0–40.0 3.6
5.3 1–37 7.0
16.0 0–65 13.0
3.4 0–65 6.4
* p < 0.001.
J Neurosurg: Pediatrics / June 6, 2014
3
C. M. Bonfield et al. TABLE 3: Mechanism of injury* Treatment Group (%) Injury Mechanism†
Non-Op
Repair
TBI
All
ATV bicycle accident fall gunshot wound lawnmower accident motorbike accident MVC NAT object to head pedestrian vs MVC rollerblading/skateboarding accident sledding accident sports collision total
23 (3.0) 18 (2.3) 395 (51.2) 1 (0.1) 0 (0.0) 9 (1.2) 85 (11.0) 33 (4.3) 130 (16.8) 39 (5.1) 12 (1.6) 0 (0.0) 27 (3.5) 772 (100)
2 (3.4) 5 (8.6) 8 (13.8) 0 (0.0) 1 (1.7) 1 (1.7) 6 (10.3) 0 (0.0) 28 (48.2) 3 (5.2) 0 (0.0) 2 (3.4) 2 (3.4) 58 (100)
4 (6.0) 4 (6.0) 12 (17.9) 1 (1.5) 0 (0.0) 1 (1.5) 16 (23.9) 6 (9.0) 12 (17.9) 8 (11.9) 2 (3.0) 0 (0.0) 1 (1.5) 67 (100)
29 (3.2) 27 (3.0) 415 (46.3) 2 (0.2) 1 (0.1) 11 (1.2) 107 (11.9) 39 (4.3) 170 (19.0) 50 (5.6) 14 (1.6) 2 (0.2) 30 (3.3) 897 (100)
* ATV = all-terrain vehicle; NAT = nonaccidental trauma. † p < 0.001 for all mechanisms between groups.
hematoma evacuation (17.9%). Other procedures in the TBI group were intraparenchymal hematoma evacuation (7.5%), penetrating injury exploration and repair (4.5%), subdural hematoma evacuation (3.0%), and posterior fossa hematoma evacuation (1.5%). Outcomes
The length of follow-up was similar in the two surgical groups (Repair 1.1 ± 1.8 years, TBI 1.0 ± 1.8 years). Complications in the surgical groups are recorded in Table 8. In the Repair group, 21 patients (36.2%) had a recorded complication. Eight patients (38.0%) had complications related to the surgical procedure, such as wound TABLE 4: Mechanism of injury and patients requiring operative intervention Injury Mechanism
No. of Cases*
ATV bicycle accident fall gunshot wound lawnmower accident motorbike accident MVC NAT object to head pedestrian vs MVC rollerblading/skateboarding accident sledding accident sports collision
6 (20.7) 9 (33.3) 20 (4.8) 1 (50.0) 1 (100.0) 2 (18.2) 22 (20.6) 6 (15.4) 40 (23.5) 11 (22.0) 2 (14.3) 2 (100.0) 3 (10.0)
* Parenthetical values are the percentage of the total of the respective mechanism.
4
infection or painful hardware necessitating removal. Thirteen patients (62.0%) had complications related to the trauma, such as new headaches or learning disabilities. Of note, no patient in the Repair group had worsening of any neurological condition (weakness, visual loss, hearing loss, and so on) after repair of the fracture. Similarly, 32 patients (42.8%) in the TBI group had complications: 3 (9.4%) were surgery related, and included cranioplasty resorption and infection, and 29 (90.6%) were a result of the trauma, such as death, hydrocephalus, or weakness.
Discussion
Head trauma is very common in the pediatric population and 10%–30% of the injuries result in skull fractures. As in our series, there is a male predominance in most reports on skull fracture.1,3,9,10,16 The most common cause of injury varies among reports, but mainly lists MVCs,1,10,15 falls,3,9,13 and assaults16 as the predominant mechanisms. However, many of these studies do not differentiate between the age of the patients or those who received surgical treatment for the fracture. In their large series on surgical management of depressed skull fractures in children, Erşahin et al. found falls and traffic accidents to be the most common causes of injury.3 In our study, although a fall was the most common injury in patients overall, those who most frequently had their fracture repaired were hit in the head by an object, and patients who required treatment for their trauma were mostly involved in an MVC. High-impact injuries cause more force to the brain, so it is not surprising that these mechanisms, like MVC, lead to more traumatic injuries. Equally, getting hit in the head directly with an object, such as a baseball, a golf club, a tree branch, or a brick, has a high potential to cause a depressed or open skull fracture. These are the fracture types that routinely undergo surgical repair. J Neurosurg: Pediatrics / June 6, 2014
Surgery for pediatric skull fractures TABLE 5: Fracture location Treatment Group (%) Fracture Location*
Non-Op
Repair
TBI
All
frontal bone temporal bone parietal bone occipital bone 2 bones 3 bones
188 (24.4) 129 (16.7) 251 (32.5) 108 (14.0) 85 (11.0) 11 (1.4)
36 (62.1) 5 (8.6) 8 (13.8) 1 (1.7) 8 (13.8) 0 (0.0)
18 (26.9) 11 (16.4) 10 (14.9) 7 (10.4) 16 (23.9) 5 (7.5)
242 (27.0) 145 (16.2) 269 (30.0) 116 (12.9) 109 (12.2) 16 (1.8)
* p < 0.001 for all fracture types between groups.
It is important to note that most patients in our large series did not require surgical intervention for their skull fracture. At birth, the skull is at 25% of its growth potential, which expands rapidly to 75% by age 2 and to 95% by the age of 10.4,14 The rapid expansion and growth allow for a greater healing and molding capacity in this population. However, skull fractures, specifically those involving the anterior cranial fossa base can also have implications for orbital development. The growth of the upper face is secondary to cerebral and ocular growth, which is not completed until the age of 6–8 years, as well as the frontal sinus, which begins the process of aeration around age 4–5 years and progresses until puberty.4 It is this lack of an aerated frontal sinus that makes it possible to treat frontal bone fractures conservatively in the young patient, unlike the teenage or adult patient. Intervention should be performed with the goals of improved cosmesis, decreased infection, and improvement of neurological defect that is the result of a depressed bone fragment or underlying hematoma. The location of the fracture on the skull is also important in determining the fracture treatment strategy. In concordance with past reports, the parietal bone was the most common bone fractured overall and in the NonOp group. However, individuals sustaining frontal bone fractures were more likely to require surgical intervention, whether for skull fracture repair or the treatment of the underlying trauma. Frontal bone fractures are more likely to involve the frontal sinus, skull base, and orbit and to have an increased chance of causing a CSF leak, ocular complications, and cosmetic deformity of the forehead. Therefore, especially in older children with aerated TABLE 6: Fracture location and patients requiring operative intervention Fracture Location
No. of Cases*
frontal bone temporal bone parietal bone occipital bone 2 bones 3 bones
54 (22.3) 16 (11.0) 18 (6.7) 8 (6.9) 24 (22.0) 5 (31.3)
* Parenthetical values are the percentage of the total of the respective mechanism.
J Neurosurg: Pediatrics / June 6, 2014
frontal sinuses, surgical repair is common. Parietal bone lesions are generally covered by hair and have a good chance of at least partially being remodeled without intervention. Having 2 or more bones involved in fractures is a factor associated with needing surgery, as well. Likely, the increased rate of intervention is a result of a more widespread fracture causing a larger cosmetic defect or a greater area in need of repair. Also, having multiple bones injured is a harbinger of a more serious and forceful traumatic event to the head, causing more damage to the skull and underlying brain. The majority of complications were a result of the trauma, rather than the surgical procedure. The rate of posttraumatic epilepsy has not been shown to decrease after elevation of a fracture.1,6 Posttraumatic headaches, learning disabilities, and behavioral problems are unlikely to be caused or improved by surgical intervention. Ten patients did die during their hospital admission as a result of the severe brain trauma; however, no deaths occurred after the initial hospitalization in any group. Only 1 patient (0.1%), initially treated conservatively, developed a growing skull fracture requiring a delayed repair. In this case, the patient was an 8-year-old boy who suffered a linear, nondisplaced fracture of his frontal bone that extended into the orbit. He returned to the outpatient clinic 8 months after the injury with a noticeable defect in his frontal bone and swelling over the site. A CT scan revealed a growing skull fracture and leptomeningeal cyst. He underwent open repair of his frontal fracture and orbital rim. Approximately 22 months later, he presented with CSF rhinorrhea, which was repaired through an endonasal endoscopic approach and a nasoseptal flap. The patient required no further intervention. To prevent these growing fractures, Sanford recommends surgically exploring wide fractures in children in whom imaging demonstrates brain herniation through the dura mater.11 Wound infection and complaints of painful hardware did occur, and must be discussed when counseling the patient and family regarding the management options. The complications requiring hardware removal all occurred in patients in whom titanium plates were used. More recently, absorbable plates and screws have been used more, which could prevent these hardware complications. Importantly, no patients who underwent surgery for repair of the fracture only had a worsening of neurological status after the intervention. 5
C. M. Bonfield et al. TABLE 7: Associated injuries Treatment Group (%) Associated Injury
Non-Op
Repair
TBI
All
p Value
intracranial hematoma facial fracture spine ophthalmological cardiac/pulmonary orthopedic abdominal/pelvic
326 (42.2) 34 (4.4) 22 (2.9) 31 (4.0) 24 (3.1) 82 (10.6) 22 (2.9)
23 (39.7) 9 (15.5) 1 (1.7) 6 (10.3) 1 (1.7) 1 (1.7) 1 (1.7)
57 (85.1) 12 (17.9) 1 (1.5) 12 (17.9) 19 (28.3) 14 (20.9) 17 (25.4)
406 (45.3) 55 (6.1) 24 (2.7) 49 (5.5) 44 (4.9) 97 (10.8) 40 (4.5)
DOI: 10.3171/2014.5.PEDS13414 ©AANS, 2014
Pediatric skull fractures: the need for surgical intervention, characteristics, complications, and outcomes Clinical article Christopher M. Bonfield, M.D.,1 Sanjay Naran, M.D., 2 Oluwaseun A. Adetayo, M.D., 2 Ian F. Pollack, M.D.,1 and Joseph E. Losee, M.D. 2 Departments of 1Neurological Surgery and 2Plastic Surgery, University of Pittsburgh, Pennsylvania Object. Head trauma is a common cause of morbidity and mortality in the pediatric population and often results in a skull fracture. Pediatric skull fractures are distinct from adult fractures. Pediatric fractures have a greater capacity to remodel, but the pediatric brain and craniofacial skeleton are still developing. Although pediatric head trauma has been extensively studied, there is sparse literature regarding skull fractures. The authors’ aim was to investigate the characteristics, injuries, complications, and outcomes of the patients in whom surgical intervention was needed for skull fractures. Methods. The authors performed a retrospective review of patients presenting to the emergency department of a pediatric Level I trauma center between 2000 and 2005 with skull fractures. Patient demographics, mechanism of injury, associated injuries, fracture bone involvement, surgical intervention, complications, and outcomes were analyzed. Groups treated nonoperatively, for skull fracture repair, and for traumatic brain injury were compared. Results. A total of 897 patients with a skull fracture were analyzed. Most patients (n = 772, 86.1%) were treated nonoperatively (Non-Op group). Fifty-eight patients (6.5%) underwent repair of the fracture (Repair group) and 67 (7.5%) required intervention for treatment of traumatic brain injury (TBI group). The Non-Op group was significantly younger, and the TBI group had a lower initial Glasgow Coma Scale (GCS) score. A fall (51.2%) was the most common mechanism of injury in the Non-Op group, whereas a motor vehicle crash (23.9%) and being hit in the head with an object (48.2%) were most prevalent in the TBI and Repair groups, respectively. Associated injuries were seen in all 3 groups, with brain injury (hematoma) being the most common. Frontal bone fracture was seen most in the Repair and TBI groups, and the parietal bone was the most frequent bone fractured in the Non-Op group. Patients in the TBI group were much more likely to have 2 or 3 skull bones fractured. In the Repair group, 36.2% had a complication (38.0% intervention related and 62.0% trauma related), but no patient had a worsening of their neurological status. In the TBI group, 48.7% of the patients suffered a complication, the vast majority (90.6%) of which were related to the trauma. Conclusions. The majority of pediatric skull fractures can be managed conservatively. Of those requiring surgical intervention, fewer than half of the surgeries are performed solely for skull fracture repair only. Patients hit in the head with an object or involved in a motor vehicle crash are more likely to need surgical intervention either to repair the skull fracture or for TBI management, respectively. Frontal bone fractures are more likely to necessitate repair, and those patients treated for TBI have a greater incidence of 2 or 3 bones involved in the fracture. Complications occurred but most were related to underlying trauma, not the surgery. No patients who underwent intervention for repair of their skull fracture only had a worsening of their neurological status. (http://thejns.org/doi/abs/10.3171/2014.5.PEDS13414)
Key Words • pediatric • skull fracture • head injury • trauma • surgical intervention
H
ead trauma is common in the pediatric population and is an important cause of morbidity and mortality in the United States, with an estimated incidence
Abbreviations used in this paper: EVD = external ventricular drain; GCS = Glasgow Coma Scale; ICU = intensive care unit; LOS = length of stay; MVC = motor vehicle crash; TBI = traumatic brain injury.
J Neurosurg: Pediatrics / June 6, 2014
of 250 per 100,000 per year. It accounts for over 7000 deaths, 60,000 hospitalizations, and 600,000 emergency department visits annually among American children.8,12 In children, between 10%–30% of head injuries result in skull fracture,7,9 many with associated brain injury.2,13 Surgical intervention is largely performed in cases of skull fracture depression, frontal sinus involvement, and underlying mass lesion. In certain instances, especially in infants, conserva1
C. M. Bonfield et al. tive management has been advocated as treatment.5,15 Pediatric skull fractures remain distinct from their adult counterparts. Pediatric skull fractures have a greater capacity to heal and remodel, but the pediatric brain and craniofacial skeleton are also still developing, which puts the children at risk for unique complications, such as growing skull fractures. Although pediatric head trauma has been extensively studied, there is sparse literature regarding skull fractures, the rate of operative management, and outcomes following surgical intervention in this population. Our goal was to investigate the characteristics, injuries, indications for intervention, and outcomes of the pediatric patients in whom surgical intervention was needed for cranial vault skull fractures.
Methods
The study was approved by the Institutional Review Board at the University of Pittsburgh. A database of all patients diagnosed with a skull fracture at the Children’s Hospital of Pittsburgh from 2000 to 2005 was searched. All patients diagnosed with a skull fracture were included regardless of treating specialty, treatment modality, or need for hospital admission. Patient demographics, mechanism of injury, associated injuries, fracture bone involvement, surgical intervention, complications, and outcomes were recorded. Comparison was made between individuals who were treated nonoperatively (Non-Op group), those surgically treated for skull fracture repair only (Repair group), and those surgically treated for traumatic brain injury (TBI) with or without repair of a skull fracture (TBI group). Indications for intervention in the Repair group included displaced or depressed fracture, open fracture, frontal sinus fracture, and cosmetic deformity. Patients in the TBI group had intervention for treatment of the underlying trauma, and the group included such operations as hematoma evacuation, decompressive hemicraniectomy, external ventricular drain (EVD) placement. Skull fractures may have been repaired in the TBI group, but that was not the principal indication for surgery. In general, for children younger than 2 years of age, sutures were used to affix the bone. For children older than 2 years of age, titanium plates and screws were used. Data analysis was performed using SPSS version 17 (SPSS, Inc.). Chi-square tests were used for the betweengroup comparisons of categorical variables. Since age, length of hospital stay, and Glasgow Coma Scale (GCS) score were not normally distributed, Kruskal-Wallis tests were used for the comparisons of continuous variables. Mean values presented ± SD.
Results Patient Characteristics
A total of 897 patients presented to Children’s Hospital of Pittsburgh between 2000 and 2005 with a diagnosis of a skull fracture. Of these patients, 772 (86.1%) were treated nonoperatively (Non-Op group). The remaining patients were grouped according to the indication for their surgical intervention. Fifty-eight patients (6.5%) un-
2
derwent repair of the fracture (Repair group) for fracture elevation, frontal sinus repair, open fracture debridement, or cosmetic repair. Sixty-seven patients (7.5%) required intervention for treatment of a TBI (TBI group) including hematoma evacuation, EVD placement, or decompressive craniectomy. Patient demographics are listed in Table 1. In all 3 groups, the majority of patients were male (63.5% in the Non-Op, 69.0% in the Repair, and 65.7% in the TBI group) and Caucasian (Non-Op 82.3%, Repair 81.0%, and TBI 85.1%). The mean age at the time of evaluation was significantly younger in the Non-Op group (5.5 ± 5.0 years, p < 0.001) compared with the Repair (8.5 ± 4.8 years) and TBI (8.6 ± 5.2) groups. Hospital Stay Data
Hospital stay data are also reported in Table 2. The initial GCS score was significantly lower in the TBI group (6.80 ± 4.5, p < 0.001) in relation to the Non-Op (13.6 ± 3.4) and Repair (12.4 ± 4.6) groups. In the TBI group, 63 patients (94.0%) were admitted to the intensive care unit (ICU) and 42 patients (62.7%) were intubated. In the Repair group, 24 patients (41.4%) were admitted to the ICU and 12 patients (20.7%) were intubated. The Non-Op group had the lowest rates of ICU admissions (19.4%) and intubations (10.0%). The length of stay (LOS) also varied between the groups. The TBI group had the longest LOS (16.0 ± 13.0 days, p < 0.001), followed by the Repair and Non-Op groups (5.3 ± 7.0 days and 2.1 ± 3.6 days, respectively).
Mechanism of Injury
The mechanisms of injury varied among the groups (Table 3). The most common injuries sustained by the Non-Op group were fall (n = 395, 51.2%), object to head (n = 130, 16.8%), and motor vehicle crash (MVC; n = 85, 11.0%). In the Repair group, the most common mechanisms of injury were object to head (n = 28, 48.2%), fall (n = 8, 13.8%), and MVC (n = 6, 10.3%). The most common injuries in the TBI group were MVC (n = 16, 23.9%), fall (n = 12, 17.9%), object to head (n = 12, 17.9%), and pedestrian hit by motor vehicle (n = 8, 11.9%). The mechanisms resulting in the highest percentage of surgical intervention (Table 4) were bicycle accident (33.3% of total bicycle accidents), object to head (23.5% of total objects to head), and pedestrian hit by motor vehicle (22.0% of total pedestrians hit by motor vehicle).
Fracture Location
Location of the fractures was also recorded (Table 5). Overall, the most common location was the parietal bone (n = 269, 30.0%), followed by the frontal bone (n = 242, 27.0%). However, the most common bone fractured differed among groups (p < 0.001). The parietal bone was the most common fracture location in the Non-Op group (n = 251, 32.5%), whereas the frontal bone was the most injured in the Repair and TBI groups (n = 36, 62.1%, and n = 18, 26.9%, respectively). In the repair group, 8 (22.2%) of 36 required surgery for frontal sinus repair and 28 (77.8%) of 36 had treatment for elevation of a deJ Neurosurg: Pediatrics / June 6, 2014
Surgery for pediatric skull fractures TABLE 1: Summary of patient demographics Treatment Group (%) Variable sex* male female total race† white black other/unknown age (yrs)‡ mean range SD
Non-Op
Repair
TBI
All
490 (63.4) 282 (36.5) 772 (100)
40 (69.0) 18 (31.0) 58 (100)
44 (65.7) 23 (34.3) 67 (100)
574 (64.0) 323 (36.0) 897 (100)
635 (82.3) 86 (11.1) 51 (6.7)
47 (81.0) 8 (13.8) 3 (5.2)
57 (85.1) 6 (9.0) 4 (6.0)
739 (82.4) 100 (11.1) 58 (6.5)
5.5 0.1–21.7 5.0
8.5 0.1–17.7 4.8
8.6 0.1–17.6 5.2
5.9 0.1–21.7 5.1
* p = 0.67. † p = 0.92. ‡ p < 0.001.
pressed fracture and cosmetic indications. The 18 patients in the TBI group had underlying brain injuries that necessitated intervention. The TBI group had the highest rate of injuries involving 2 (23.9%) and 3 (7.5%) bones. The Non-Op group (2 bones in 11.0% and 3 bones in 1.4%) and the Repair group (2 bones in 13.8% and 3 in bones 0.0%) had lower rates of multiple bone involvement in fractures. The fractured bones most likely to result in a patient undergoing surgical treatment (Table 6) were 3 bones involved (31.3% of total 3 bone–involved injury), frontal bone (22.3% of total frontal bone injury) and 2 bones involved (22% of total 2 bone–involved injury). Associated Injuries
Many patients sustained injuries to other areas of body besides the skull. Associated injury rates are reported in Table 7. Intracranial hemorrhage (epidural, subdural, subarachnoid, or intraparenchymal) was the
most common associated injury in all 3 groups (Non-Op group: n = 326, 42.2%; Repair group: n = 23, 39.7%; and TBI group: n = 57, 85.1%). The next most encountered associated injury was orthopedic (n = 82, 10.6%) in the Non-Op group, face (n = 9, 15.5%) in the Repair group, and cardiac/pulmonary (n = 19, 28.3%) in the TBI group. Procedures
The types of intervention in the surgical groups were reviewed. All fractures, except one, in the Repair group were depressed (n = 57, 98.3%), and 21 fractures (36.2%) were open. These fractures underwent procedures for elevation, repair, or debridement. The remaining patient underwent the repair of a growing skull fracture 8 months after the initial injury. In the TBI group, the most common procedure was the insertion of an EVD (77.6%). The most common open surgical intervention was a decompressive craniectomy (22.4%), followed by epidural
TABLE 2: Hospital stay data Treatment Group (%) Variable GCS score on admission* mean range SD ICU admission* intubated* LOS (days)* mean range SD
Non-Op
Repair
TBI
All
13.6 3–15 3.4 150 (19.4) 77 (10.0)
12.4 3–15 4.6 24 (41.4) 12 (20.7)
6.8 3–15 4.5 63 (94.0) 42 (62.7)
13.0 3–15 4.0 237 (26.4) 131 (14.6)
2.1 0–40.0 3.6
5.3 1–37 7.0
16.0 0–65 13.0
3.4 0–65 6.4
* p < 0.001.
J Neurosurg: Pediatrics / June 6, 2014
3
C. M. Bonfield et al. TABLE 3: Mechanism of injury* Treatment Group (%) Injury Mechanism†
Non-Op
Repair
TBI
All
ATV bicycle accident fall gunshot wound lawnmower accident motorbike accident MVC NAT object to head pedestrian vs MVC rollerblading/skateboarding accident sledding accident sports collision total
23 (3.0) 18 (2.3) 395 (51.2) 1 (0.1) 0 (0.0) 9 (1.2) 85 (11.0) 33 (4.3) 130 (16.8) 39 (5.1) 12 (1.6) 0 (0.0) 27 (3.5) 772 (100)
2 (3.4) 5 (8.6) 8 (13.8) 0 (0.0) 1 (1.7) 1 (1.7) 6 (10.3) 0 (0.0) 28 (48.2) 3 (5.2) 0 (0.0) 2 (3.4) 2 (3.4) 58 (100)
4 (6.0) 4 (6.0) 12 (17.9) 1 (1.5) 0 (0.0) 1 (1.5) 16 (23.9) 6 (9.0) 12 (17.9) 8 (11.9) 2 (3.0) 0 (0.0) 1 (1.5) 67 (100)
29 (3.2) 27 (3.0) 415 (46.3) 2 (0.2) 1 (0.1) 11 (1.2) 107 (11.9) 39 (4.3) 170 (19.0) 50 (5.6) 14 (1.6) 2 (0.2) 30 (3.3) 897 (100)
* ATV = all-terrain vehicle; NAT = nonaccidental trauma. † p < 0.001 for all mechanisms between groups.
hematoma evacuation (17.9%). Other procedures in the TBI group were intraparenchymal hematoma evacuation (7.5%), penetrating injury exploration and repair (4.5%), subdural hematoma evacuation (3.0%), and posterior fossa hematoma evacuation (1.5%). Outcomes
The length of follow-up was similar in the two surgical groups (Repair 1.1 ± 1.8 years, TBI 1.0 ± 1.8 years). Complications in the surgical groups are recorded in Table 8. In the Repair group, 21 patients (36.2%) had a recorded complication. Eight patients (38.0%) had complications related to the surgical procedure, such as wound TABLE 4: Mechanism of injury and patients requiring operative intervention Injury Mechanism
No. of Cases*
ATV bicycle accident fall gunshot wound lawnmower accident motorbike accident MVC NAT object to head pedestrian vs MVC rollerblading/skateboarding accident sledding accident sports collision
6 (20.7) 9 (33.3) 20 (4.8) 1 (50.0) 1 (100.0) 2 (18.2) 22 (20.6) 6 (15.4) 40 (23.5) 11 (22.0) 2 (14.3) 2 (100.0) 3 (10.0)
* Parenthetical values are the percentage of the total of the respective mechanism.
4
infection or painful hardware necessitating removal. Thirteen patients (62.0%) had complications related to the trauma, such as new headaches or learning disabilities. Of note, no patient in the Repair group had worsening of any neurological condition (weakness, visual loss, hearing loss, and so on) after repair of the fracture. Similarly, 32 patients (42.8%) in the TBI group had complications: 3 (9.4%) were surgery related, and included cranioplasty resorption and infection, and 29 (90.6%) were a result of the trauma, such as death, hydrocephalus, or weakness.
Discussion
Head trauma is very common in the pediatric population and 10%–30% of the injuries result in skull fractures. As in our series, there is a male predominance in most reports on skull fracture.1,3,9,10,16 The most common cause of injury varies among reports, but mainly lists MVCs,1,10,15 falls,3,9,13 and assaults16 as the predominant mechanisms. However, many of these studies do not differentiate between the age of the patients or those who received surgical treatment for the fracture. In their large series on surgical management of depressed skull fractures in children, Erşahin et al. found falls and traffic accidents to be the most common causes of injury.3 In our study, although a fall was the most common injury in patients overall, those who most frequently had their fracture repaired were hit in the head by an object, and patients who required treatment for their trauma were mostly involved in an MVC. High-impact injuries cause more force to the brain, so it is not surprising that these mechanisms, like MVC, lead to more traumatic injuries. Equally, getting hit in the head directly with an object, such as a baseball, a golf club, a tree branch, or a brick, has a high potential to cause a depressed or open skull fracture. These are the fracture types that routinely undergo surgical repair. J Neurosurg: Pediatrics / June 6, 2014
Surgery for pediatric skull fractures TABLE 5: Fracture location Treatment Group (%) Fracture Location*
Non-Op
Repair
TBI
All
frontal bone temporal bone parietal bone occipital bone 2 bones 3 bones
188 (24.4) 129 (16.7) 251 (32.5) 108 (14.0) 85 (11.0) 11 (1.4)
36 (62.1) 5 (8.6) 8 (13.8) 1 (1.7) 8 (13.8) 0 (0.0)
18 (26.9) 11 (16.4) 10 (14.9) 7 (10.4) 16 (23.9) 5 (7.5)
242 (27.0) 145 (16.2) 269 (30.0) 116 (12.9) 109 (12.2) 16 (1.8)
* p < 0.001 for all fracture types between groups.
It is important to note that most patients in our large series did not require surgical intervention for their skull fracture. At birth, the skull is at 25% of its growth potential, which expands rapidly to 75% by age 2 and to 95% by the age of 10.4,14 The rapid expansion and growth allow for a greater healing and molding capacity in this population. However, skull fractures, specifically those involving the anterior cranial fossa base can also have implications for orbital development. The growth of the upper face is secondary to cerebral and ocular growth, which is not completed until the age of 6–8 years, as well as the frontal sinus, which begins the process of aeration around age 4–5 years and progresses until puberty.4 It is this lack of an aerated frontal sinus that makes it possible to treat frontal bone fractures conservatively in the young patient, unlike the teenage or adult patient. Intervention should be performed with the goals of improved cosmesis, decreased infection, and improvement of neurological defect that is the result of a depressed bone fragment or underlying hematoma. The location of the fracture on the skull is also important in determining the fracture treatment strategy. In concordance with past reports, the parietal bone was the most common bone fractured overall and in the NonOp group. However, individuals sustaining frontal bone fractures were more likely to require surgical intervention, whether for skull fracture repair or the treatment of the underlying trauma. Frontal bone fractures are more likely to involve the frontal sinus, skull base, and orbit and to have an increased chance of causing a CSF leak, ocular complications, and cosmetic deformity of the forehead. Therefore, especially in older children with aerated TABLE 6: Fracture location and patients requiring operative intervention Fracture Location
No. of Cases*
frontal bone temporal bone parietal bone occipital bone 2 bones 3 bones
54 (22.3) 16 (11.0) 18 (6.7) 8 (6.9) 24 (22.0) 5 (31.3)
* Parenthetical values are the percentage of the total of the respective mechanism.
J Neurosurg: Pediatrics / June 6, 2014
frontal sinuses, surgical repair is common. Parietal bone lesions are generally covered by hair and have a good chance of at least partially being remodeled without intervention. Having 2 or more bones involved in fractures is a factor associated with needing surgery, as well. Likely, the increased rate of intervention is a result of a more widespread fracture causing a larger cosmetic defect or a greater area in need of repair. Also, having multiple bones injured is a harbinger of a more serious and forceful traumatic event to the head, causing more damage to the skull and underlying brain. The majority of complications were a result of the trauma, rather than the surgical procedure. The rate of posttraumatic epilepsy has not been shown to decrease after elevation of a fracture.1,6 Posttraumatic headaches, learning disabilities, and behavioral problems are unlikely to be caused or improved by surgical intervention. Ten patients did die during their hospital admission as a result of the severe brain trauma; however, no deaths occurred after the initial hospitalization in any group. Only 1 patient (0.1%), initially treated conservatively, developed a growing skull fracture requiring a delayed repair. In this case, the patient was an 8-year-old boy who suffered a linear, nondisplaced fracture of his frontal bone that extended into the orbit. He returned to the outpatient clinic 8 months after the injury with a noticeable defect in his frontal bone and swelling over the site. A CT scan revealed a growing skull fracture and leptomeningeal cyst. He underwent open repair of his frontal fracture and orbital rim. Approximately 22 months later, he presented with CSF rhinorrhea, which was repaired through an endonasal endoscopic approach and a nasoseptal flap. The patient required no further intervention. To prevent these growing fractures, Sanford recommends surgically exploring wide fractures in children in whom imaging demonstrates brain herniation through the dura mater.11 Wound infection and complaints of painful hardware did occur, and must be discussed when counseling the patient and family regarding the management options. The complications requiring hardware removal all occurred in patients in whom titanium plates were used. More recently, absorbable plates and screws have been used more, which could prevent these hardware complications. Importantly, no patients who underwent surgery for repair of the fracture only had a worsening of neurological status after the intervention. 5
C. M. Bonfield et al. TABLE 7: Associated injuries Treatment Group (%) Associated Injury
Non-Op
Repair
TBI
All
p Value
intracranial hematoma facial fracture spine ophthalmological cardiac/pulmonary orthopedic abdominal/pelvic
326 (42.2) 34 (4.4) 22 (2.9) 31 (4.0) 24 (3.1) 82 (10.6) 22 (2.9)
23 (39.7) 9 (15.5) 1 (1.7) 6 (10.3) 1 (1.7) 1 (1.7) 1 (1.7)
57 (85.1) 12 (17.9) 1 (1.5) 12 (17.9) 19 (28.3) 14 (20.9) 17 (25.4)
406 (45.3) 55 (6.1) 24 (2.7) 49 (5.5) 44 (4.9) 97 (10.8) 40 (4.5)
