ORIGINAL ARTICLE
Basal skull fractures are associated with mortality in pediatric severe traumatic brain injury Ibrahim Alhelali, MD, Tanya Charyk Stewart, MSc, Jennifer Foster, MD, Ibrahim M. Alharfi, MD, Adrianna Ranger, MD, Hani Daoud, MD, and Douglas D. Fraser, MD, PhD, London, Ontario, Canada
Basal skull fractures (BSFs) are caused by blunt force trauma, occurring in the temporal, occipital, sphenoid, and/or ethmoid bones. In pediatric severe traumatic brain injury (sTBI), there is a paucity of data on BSFs. Our goal was to investigate the BSF prevalence, anatomy, and association with short-term outcomes in pediatric sTBI. METHODS: We retrospectively reviewed all severely injured (Injury Severity Score Q12) pediatric patients (aged G18 years) admitted to our hospital after experiencing an sTBI (Glasgow Coma Scale score e8 and head Abbreviated Injury Scale score Q4). Neuroimaging for all sTBI patients was reviewed for skull fractures. Data were analyzed with both univariate and multivariate techniques. RESULTS: Of the 180 patients with sTBI, 47 had BSFs for a prevalence of 26% (69 BSFs in total; 16 sTBI patients had Q2 BSFs). The squamous temporal bone was fractured most frequently (n=30/47 sTBI patients with BSFs). Patients with BSFs were heavier and had more facial injuries than those without (p G 0.05) but were similar in all other admission demographics, injury profiles, and clinical characteristics. Cerebrospinal fluid leak was found in 32% (n = 15 of 47) of BSF patients (otorrhea, n = 12; rhinorrhea, n = 1; otorrhea/rhinorrhea, n = 2; p G 0.001). Mortality, acute central diabetes insipidus, and fewer ventilator-free days were associated with BSFs (p G 0.005), whereas in sTBI survivors, BSFs were associated with longer lengths of stay (p G 0.05). Multiple logistic regression showed that BSFs were positively associated with the presence of subarachnoid hemorrhage (odds ratio [OR], 4.00; p = 0.001), contusion (OR, 2.48; p = 0.029), herniation (OR, 3.40; p = 0.037), and cerebral edema (OR, 2.30; p = 0.047) but negatively associated with diffuse axonal injury (OR, 0.20; p = 0.003). BSFs and mortality were strongly associated (OR, 6.87; p = 0.019). CONCLUSION: BSFs occurred in 26% of pediatric sTBI patients. The temporal bone was fractured in two thirds of sTBI patients with BSFs, and one third was associated with cerebrospinal fluid leaks. BSFs represent a significant linear blunt force and are independent predictors of mortality. (J Trauma Acute Care Surg. 2015;78: 1155Y1161. Copyright * 2015 Wolters Kluwer Health, Inc. All rights reserved.) LEVEL OF EVIDENCE: Prognostic and epidemiologic study, level III. KEY WORDS: Traumatic brain injury; children; adolescents; skull fracture; mortality. BACKGROUND:
BACKGROUND Severe traumatic brain injury (sTBI) is a leading cause of death and disability among children.1Y4 Outcome prediction in the pediatric sTBI patients is difficult, and outcomes may vary significantly between centers.5,6 Abnormalities found on head imaging may aid sTBI prognostication,7Y9 particularly when combined with clinical data.6,10Y12 Skull fractures are relatively common findings on head imaging, reported in 43% to 62% of pediatric trauma patients.10,13 The frequency of basal skull fractures (BSFs) in the Submitted: December 10, 2014, Revised: February 26, 2015, Accepted: February 27, 2015. From the Departments of Paediatrics (I.A., J.F., I.M.A., H.D., D.D.F.) and Surgery (T.C.S., A.R.), Western University; Trauma Program (T.C.S.), London Health Sciences Center; and Children’s Health Research Institute (J.F., D.D.F.), London, Ontario, Canada; Department of Pediatric Critical Care (I.M.A.), King Fahad Medical City, Riyadh, Saudi Arabia; and Translational Research Centre (D.D.F.); Physiology and Pharmacology (D.D.F.) and Clinical Neurological Sciences (D.D.F.), Western University; and Centre for Critical Illness Research (D.D.F.), London, Ontario, Canada. Address for reprints: Douglas D. Fraser, MD, PhD, Room C2-C82, Children’s Hospital, London Health Sciences Centre, 800 Commissioners Rd. E., London, Ontario, Canada N6A 5W9; email: [email protected]. DOI: 10.1097/TA.0000000000000662
general pediatric TBI population has been reported to be 4% to 14%7,14Y17 and in one pediatric sTBI cases series to be 20%.10 This latter publication did not evaluate BSFs specifically in outcome analyses. BSFs generally occur when a linear skull fracture with a vertical or oblique orientation extends to the base of the skull.18 The skull base consists of the basilar portion of the temporal, occipital, sphenoid, and ethmoid bones. A common clinical sign of BSF is subcutaneous bleeding occurring over the mastoid process or orbit. Of particular concern, BSFs can be associated with cranial nerve injury, vascular disruption, hearing loss, and cerebrospinal fluid (CSF) leak.19Y22 Dural tears can result from BSFs23 and may be a potential entry point for bacterial invasion and subsequent meningitis, but antibiotic prophylaxis remains controversial.24 Based on the paucity of data surrounding BSF in the pediatric sTBI population, the aims of this study were to determine the BSF prevalence, anatomic location, and their relationship to short-term outcomes.
PATIENTS AND METHODS The Health Sciences Research Ethics Board at Western University approved this study. We screened all pediatric trauma
J Trauma Acute Care Surg Volume 78, Number 6
1155
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Trauma Acute Care Surg Volume 78, Number 6
Alhelali et al.
patients (91 month and G18 years) during a 12-year period (January 2000YDecember 2011) with an Injury Severity Score (ISS) of 12 or higher. Patients with an sTBI were identified by a pre-sedation Glasgow Coma Scale (GCS) score of 8 or less and a head Maximum Abbreviated Injury Scale (MAIS) score of 4 or higher.25Y29 All sTBI patients were admitted to the pediatric intensive care unit (PICU) located in the Children’s Hospital, London Health Sciences Centre (CH-LHSC), which is a regional pediatric Level I trauma center for Southwestern Ontario. CH-LHSC serves a geographic area of 19,000 km2 with a pediatric population of more than 500,000. Patients who were admitted more than 12 hours after their injury were excluded. Patients were identified from the prospectively collected written and electronic admission records of the PICU, collated with the data of the provincially mandated LHSC Trauma Registry.25Y28 The patient data collected included demographics, injury data, neuroimaging studies, and outcomes. Pre-sedation GCS was determined at either the scene or referring hospital or on arrival to LHSC trauma center. The pupillary response and hypotension (systolic blood pressure [SBP] G70 mm Hg for infants, SBP G70 mm Hg + (2 * age) for toddlers and children younger than 10 years, or SBP G90 mm Hg for children aged 10 years or older) were recorded on arrival to our trauma room. The reports of computed tomography (CT) imaging were collected. Specific abnormalities documented by fellowshiptrained neuroradiologists included skull fractures, cerebral edema, diffuse axonal injury (DAI), subarachnoid hemorrhage (SAH), subdural hemorrhage (SDH), epidural hemorrhage (EDH), intraventricular hemorrhage (IVH), brain herniation, midline shift, cerebral contusion, and ischemia. Indications for acute CT angiography included BSFs extending to the carotid canal, the jugular foramen, and/or the foramen magnum and suspected injury to the sagittal sinus and/ or brain ischemia in a large vessel distribution. Magnetic resonance (MR) angiography and/or venography was used in follow-up or in cases of evolving ischemic lesion. During the study period, patients with sTBI in our institution were placed on a temperature-modulating blanket in supine position, with the head elevated to 30 degrees. Normothermia was targeted in those sTBI patients not enrolled in a hypothermia trial30 with administration of antipyretics, passive cooling, and/or use of the temperature-modulating blanket. Antibiotics were not routinely administered, except for the inconsistent prophylactic use of cefazolin after intracranial pressure (ICP) monitor placement.31 Adequate analgesia and sedation were obtained with opioid and benzodiazepine infusions, respectively. Raised ICP was generally managed as per published previously.25Y29 If the ICP is more than 20 mm Hg to 25 mm Hg for longer than 5 minutes or for a rapidly raising ICP: 1) Drain CSF for 5 minutes if external ventricular drain (EVD) in situ;31 2) Mannitol 0.5 g/kg intravenously for 20 minutes every 6 hours as needed; 3) 3% NaCl given in boluses of 1 mL/kg to 2 mL/kg for 5 minutes every 12 hours as needed; 4) Hyperventilation to a PaCO2 35 mm Hg; 5) Barbiturate infusion titrated to ICP and cerebral perfusion pressure; and 6) Neurosurgical consult for potential decompressive craniectomy. Mannitol and 3% NaCl were held if the measured osmolality was greater than 320 mOsm/L and 360 mOsm/L, respectively. Low cerebral perfusion pressure secondary to arterial hypotension 1156
and without raised ICP was managed as follows: 0.9% NaCl (or colloid) 10 mL/kg intravenously for 5 to 30 minutes as needed, followed by administration of inotropes/vasopressors. Severe TBI outcomes included in-hospital mortality, ventilator-free days (unventilated days in the first 28 days of admission), PICU and hospital lengths of stay (LOS), and discharge destination. Discharge destinations included chronic rehabilitation hospital, acute care hospital, or home. Discharge to a chronic rehabilitation hospital was a less favorable outcome because it indicated that the patient was less functionally independent on discharge. CSF leak was determined by the presence of BSF on head CT and documentation of a clinical CSF leak, with or without glucose testing.32 Acute central diabetes insipidus (CDI) was defined as polyuria (urine output 9 4 mL/kg per hour for children G70 kg; 9300 mL/h for adult-size children Q70 kg) for at least two consecutive hours, hypernatremia (serum Na 9145 mmol/L), high serum osmolality (9300 mOsm/kg), and low urine osmolality (G300 mOsm/kg) at the time of diagnosis.27 Confirmation of CDI was based on symptom reversal with administration of 1desamino-8-D-arginine vasopressin (DDAVP) and/or vasopressin. Infections in sTBI patients were described previously.25 Culture-positive urinary tract infections (single organism; Q105 colony-forming units/mL) were determined within the first 48 hours of admission. Nosocomial infections recorded were identified by Centers for Disease Control criteria and included ventilator-associated pneumonia,33 catheter-associated urinary tract infection,34 and central line bloodstream infection.35 Meningitis and wound infections were confirmed by positive cultures of a single organism. Continuous variables were found to deviate from normality and, hence, medians with interquartile ranges (IQRs) were presented. Between-group comparisons were made using the Mann-Whitney U-test. For categoric variables, the frequencies and percentages were presented and between-group comparisons were made using the Pearson’s W2 or Fisher’s exact test where appropriate. PICU LOS, hospital LOS and postdischarge destination were assessed for survivors only. For all analyses, a value of p G 0.05 was considered statistically significant. Multivariate logistic regression modeling was performed with BSF as the outcome variable. Variables identified a priori to be assessed for inclusion in the model were age, gender, ISS, motor vehicle collision, abnormalities identified on admission head CT (DAI, cerebral edema, contusion, ischemia, herniation, midline shift, EDH, SDH, SAH, and IVH), and the presence of a serious facial injury (defined as MAIS 3Y6 in the face). Variables found to be significant in univariate analyses at the 0.25 level were then entered and allowed to be removed from the model at the 0.10 level in a backward elimination strategy. Sensitivity analyses were performed using different modeling methods (i.e., forward stepwise logistic regression).36 A second multivariate logistic regression analysis was undertaken to determine the association of patient’s demographics, physiologic variables, neuroimaging abnormalities, including BSF, and overall injury severity on mortality, while controlling for the possible confounding effects of these variables on the relationships. Possible confounders were identified a priori. Variables considered included age (assessed both as a continuous variable and dichotomized into adolescent [aged 13Y17 years] or * 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Trauma Acute Care Surg Volume 78, Number 6
Alhelali et al.
nonadolescent [aged G12 years]), gender, ISS, hypotension on arrival to the hospital, pre-sedation motor GCS, fixed pupillary response on hospital admission, CDI, and all abnormalities identified on head CT. A backward elimination procedure was used in which variables considered a priori to be clinically important (i.e., ISS, BSF) were forced into the model. Other variables found to be significant in univariate analyses at the 0.25 level were then entered and allowed to be removed from the model at the 0.10 level. Sensitivity analyses were performed using different modeling methods (i.e., forward stepwise logistic regression).36 For both models, the Hosmer-Lemeshow statistic was calculated to evaluate the model fit and the C statistic was calculated to evaluate the predictive accuracy of our logistic regression model. Correlations among independent variables were used to assess potential multicollinearity. Variables with strong collinearity were removed, and sensitivity analyses were performed. The final models were used to determine the estimated odds ratio (OR) for either BSF or mortality, respectively, for each variable adjusted for confounders. All analyses were performed using IBM SPSS Statistics Version 21 (IBM Corporation, Armonk, NY).
RESULTS Of 818 pediatric trauma patients screened, 180 cases were identified that met study inclusion criteria for sTBI (GCS score e8 and MAIS score Q 4). All sTBI patients were admitted to the PICU. Skull fractures, both vault and BSF, were found in 39% of sTBI patients (n = 71 of 180). A total of 69 BSFs were found in the 47 sTBI patients, with the squamous temporal bone fractured most frequently (n = 30 of 47 sTBI patients; Fig. 1). Sixteen sTBI patients had two or more fractures, with the most common fracture combination being in temporal and occipital bones (n = 7 patients). CT angiography was completed in 7 of the 47 patients with BSFs. Indications for CT angiography included fractures extending to the carotid canal (n = 5), the foramen magnum (n = 2), and the jugular foramen (n = 1). MR angiography (n = 2) and MR venography (n = 1) were used in follow-up.
No vascular lesions were identified in the patient population investigated. Patients in the BSF group had greater weight on admission and median facial MAIS scores than those without BSFs (p G 0.05), but the groups were otherwise similar with respect to admission demographics, injuries, and clinical characteristics (Table 1). Patients with BSFs also had a greater number of other admission head CT abnormalities, including SAH, cerebral edema, contusion, and herniation (p G 0.005; Table 2). In contrast, DAI was more common in patients without BSFs (p G 0.05; Table 2). There were significantly higher rates of mortality and acute CDI in patients with BSFs, as well as fewer ventilatorfree days (p e 0.003; Table 3). CSF leaks occurred only in sTBI patients with BSFs (n = 15 of 47) and consisted of otorrhea (n = 12), rhinorrhea (n = 1), or otorrhea/rhinorrhea (n = 2). Two sTBI patients developed meningitis, but only one patient had a BSF. Decompressive craniectomy was performed on 28% of sTBI patients with BSFs (n = 13 of 47) as compared with 9% of sTBI patients without BSFs (n = 12 of 133; p = 0.001). For sTBI survivors, those with BSFs had greater PICU and hospital LOS (p G 0.05; Table 3). CSF leaking resolved spontaneously in all surviving patients, albeit 45% of sTBI patients with BSFs had acute placement of EVDs with intermittent CSF drainage as part of our ICP control protocol (n = 21 of 47). A multivariate logistic regression modeling analysis was conducted to examine the association of patient and injury variables with BSFs (Table 4). Variables were analyzed for all two-way interactions, and as none were found to be significant and they were removed from the model. Our sensitivity analysis revealed two nearly identical models with both forward stepwise and backward elimination and with no multicollinearity of the variables evident in either model. The strength of association for all variables was also very similar in both models. The only difference between the models was that age was included in the backward elimination, as age was not statistically associated with sustaining a BSF (OR, 1.07; p = 0.07). For the model performance evaluation, a test of the full model against
Figure 1. A schematic diagram illustrating 69 BSF locations and frequencies in 47 pediatric sTBI patients (16 patients had Q2 BSFs). * 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
1157
J Trauma Acute Care Surg Volume 78, Number 6
Alhelali et al.
TABLE 1. Admission Demographics, Injuries, and Clinical Criteria for sTBI Patients With and Without BSF Variable
BSF (n = 47) No BSF (n = 133)
Age, y Weight, kg* Male, n (%) Etiology, n (%) MVC Fall Assault/abuse Other Injury Severity Score Injury profile MAIS head MAIS neck MAIS face* MAIS chest MAIS abdomen MAIS extremity/pelvis MAIS external Presedation GCS score Presedation motor GCS score Absent pupillary response, n (%) Hypotension, n (%)
15 (8) 60 (55)* 32 (68)
12 (12) 45 (48)* 81 (61)
TABLE 3. Short-term Outcomes in sTBI Patients With and Without BSF p
0.235 0.038* 0.381 0.471
35 (75) 5 (11) 4 (9) 3 (6) 34 (17)
100 (75) 12 (9) 18 (14) 3 (2) 35 (17)
0.657
5 (0) 2.5 (3) 2 (2)* 3 (1) 2.5 (1) 3 (1) 1 (0) 4 (4) 2 (3) 18 (38) 16 (34)
5 (0) 2 (0) 1 (1)* 4 (1) 2 (2) 3 (1) 1 (0) 5 (4) 2 (3) 36 (27) 27 (20)
0.554 0.352 0.018* 0.347 0.686 0.174 0.422 0.218 0.352 0.149 0.058
Continuous variables are presented as median (IQR) and categoric variables as n (%). MVC, motor vehicle collision.
a constant-only model was statistically significant, indicating that the predictor variables reliably distinguished between patients with or without BSFs (Omnibus Test of model coefficients, X62 = 43.59, p G 0.001). There was no evidence of a lack of fit (Hosmer and Lemeshow X72 = 7.05, p = 0.424), with an overall classification success of BSF at 80% for the final model. The C statistic was 0.809, indicating good predictive accuracy of our model. SAH (OR, 4.00; p = 0.001), contusion (OR, 2.48; p = 0.029), herniation (OR, 3.40; p = 0.037), and cerebral edema (OR, 2.30; p = 0.047) were found to be positively associated with sustaining a BSF. In contrast, DAI was negatively associated with BSF (OR, 0.20; p = 0.003).
TABLE 2. Admission Head CT Abnormalities in sTBI Patients With and Without BSF (n = 171*) Head CT Abnormalities SAH, n (%)† Cerebral edema, n (%)† Contusion, n (%)† Herniation, n (%)† DAI, n (%)† Midline shift, n (%) SDH, n (%) IVH, n (%) Ischemia, n (%) EDH, n (%)
BSF (n = 47)
No BSF (n = 124)
p
29 (62)† 31 (66)† 23 (49)† 10 (21)† 6 (13)† 12 (26) 25 (53) 8 (17) 7 (15) 6 (13)
42 (34)† 49 (40)† 32 (26)† 7 (6)† 37 (30)† 19 (15) 53 (43) 32 (26) 10 (8) 11 (8)
0.001† 0.002† 0.004† 0.004† 0.022† 0.122 0.221 0.226 0.250 0.391
*Only 171 admission head CTs were available for review.
1158
Outcomes
BSF (n = 47) No BSF (n = 133)
Mortality, n (%)* 21 (45)* Central diabetes insipidus, n (%)* 15 (32)* Cerebrospinal fluid leak, n (%)* 15 (32)* Ventilator-free days, n (%)* 14 (21)* Infection, n (%)† 8 (17) For survivors only* n = 26* Hospital LOS* 30.5 (26)* PICU LOS* 10.5 (8)* Discharge to, n (%) Home 17 (65) Acute care 1 (4) Rehabilitation 8 (31)
29 (22)* 17 (13)* 0 (0)* 23 (18)* 22 (17) n = 104* 18 (27)* 5 (8)*
p 0.003* 0.003* G0.001* G0.001* 0.940 0.046* 0.018* 0.421
68 (65) 12 (12) 24 (23)
†Two patients had meningitis, but only one patient had a BSF of the temporal bone without CSF leak. Continuous variables are presented as median (IQR) and categoric variables as n (%).
A second multivariate logistic regression modeling analysis was conducted to examine the association of patient and injury variables, including the effect of sustaining a BSF, on mortality (Table 5). Variables were analyzed for two-way interactions, and as none were found to be significant, they were removed from the model. Acute CDI was correlated with GCS motor, so acute CDI was removed. There was no other multicollinearity between the other variables in the model. We also evaluated cranial vault fractures; however, they were correlated with BSFs. Because cranial vault fractures had the weaker strength of association with mortality and a higher p value, they were removed. Sensitivity analysis confirmed the effect of the different types of skull fractures on mortality. Logistic regression models that included (i) cranial vault fractures, (ii) BSFs, or (iii) all skull fractures showed that only the BSFs were statistically associated with mortality and, thus, only BSFs were included in the final model. The unadjusted OR of BSF on death was 2.90 (95% confidence interval [95% CI], 1.43Y5.88; p = 0.003). A test of the full model against a constant-only model was statistically significant, indicating that the set of predictor variables reliably distinguished between survivors and nonsurvivors (Omnibus Test of model coefficients, X72 = 135.002, p G 0.001). The Hosmer and
TABLE 4. Multivariate Logistic Regression Model With BSF as the Outcome Variable Variable SAH* DAI* Contusion* Herniation* Cerebral edema* Hypotension
A
SE
OR
95% CI
p
1.39* j1.64* 0.91* 1.22* 0.83* 0.86
0.42* 0.55* 0.42* 0.59* 0.42* 0.45
4.00* 0.20* 2.48* 3.40* 2.30* 2.36
1.75Y9.15* 0.07Y0.57* 1.10Y5.60* 1.08Y10.73* 1.01Y5.21* 0.97Y5.69
0.001* 0.003* 0.029* 0.037* 0.047* 0.057
Constant A = j2.43; SE = 0.42.
* 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Trauma Acute Care Surg Volume 78, Number 6
Alhelali et al.
TABLE 5. Multivariate Logistic Regression Model With Death as the Outcome Variable Variable
A
SE
OR
95% CI
p
Absent pupillary response* 3.42* 0.72* 30.47* 7.37Y126.00* G0.001* GCS motor* j1.58* 0.48* 0.21* 0.08Y0.53* 0.001* BSF* 1.93* 0.82* 6.87* 1.38Y34.34* 0.019* Hypotension* 1.64* 0.77* 5.14* 1.13Y23.34* 0.034* Ischemia* 2.15* 1.09* 8.58* 1.02Y72.13* 0.048* Injury Severity Score 0.06 0.04 1.06 0.995Y1.14 0.069 Cerebral edema 1.27 0.71 3.56 0.89Y14.27 0.073 Constant A = j4.11; SE = 1.61.
Lemeshow test indicated a good fit of the model (X82 = 6.510; p = 0.590), with overall classification success of death at 90%. The C statistic was 0.967, indicating excellent predictive accuracy of our model. The Wald criterion demonstrated that mortality was independently associated with BSF (OR, 6.87; p = 0.019), along with other variables, including absent pupillary response (OR, 30.47; p G 0.001), GCS motor (OR, 0.21; p = 0.001), hypotension on arrival (OR, 5.14; p = 0.034), and ischemia on admission head imaging (OR, 8.58; p = 0.048).
DISCUSSION In this study, we report the prevalence of skull fractures in a cohort of well-described pediatric sTBI patients.25Y29 Skull fractures were found in 39% of sTBI patients, with BSFs observed in 26%. Two or more BSFs were found in 9% of sTBI patients. About two thirds of BSFs occurred in the temporal bone, and one third of BSFs were associated with CSF leaks. BSF was positively associated with admission head CT abnormalities including SAH, contusion, herniation, and cerebral edema but negatively associated with DAI. Mortality was strongly associated with BSFs. To our knowledge, this is the first study to fully characterize a population of pediatric sTBI patients with BSFs. The prevalence of BSFs in patients with sTBI in this cohort was 26%, consistent with a previous case series.10 The prevalence of BSFs in the general pediatric trauma population is much lower at 4% to 14%,7,14Y17 albeit difficult to accurately ascertain because clinical features are not reliable indicators of their presence and many patients with mild TBI will not undergo a CT scan. Linear blunt force trauma causes BSFs. Our data indicate that BSFs were associated with SAH, contusion, cerebral edema, and herniation. In contrast, DAI was negatively associated with BSF, a finding that is consistent with angular, or rotational, head injuries.37 The majority of BSFs in our study were in the temporal bone, an area of the skull considered least capable of absorbing a blow and, thus, the most susceptible to fracture.38 The distribution of fracture locations was similar to that seen in other reports.19,39 Similarly, the gender distribution (approximately two-thirds male) for the entire sTBI population and those with BSF in particular is consistent with the trauma literature.6,40Y42 Patients with BSFs had higher MAIS face,
suggesting that those trauma patients with serious facial injuries should be screened for BSFs. BSFs are associated with vascular disruption, CSF leak, and meningitis.19Y22 We did not identify any sTBI patients with large vessel occlusion/dissection, a finding consistent with the rarity of large vessel injury in pediatric BSFs; carotid injury associated with BSFs were estimated at 0.09%.43 CSF leaks were found in approximately one third of sTBI patients with BSFs, with the majority of cases otorrhea. In sTBI survivors, all CSF leaks resolved spontaneously, which is likely a factor of anatomy, as otorrhea secondary to temporal bone fracture tends to resolve spontaneously.44 Leak resolution also may have been hastened by our frequent use of acute EVD drainage.45 Meningitis developed in two sTBI patients, but only one of these patients had a BSF. Hence, our data were unable to establish any relationships between BSF, CSF leaks, and infection. Meningitis could occur secondary to bacterial seeding, dural tears, and/or from placement of an EVD.31 In our sTBI cohort, patients with BSFs had a higher occurrence of SAH, cerebral edema, and herniation, all indicating a higher severity of injury than in those sTBI patients without BSFs. Virtually all clinical short-term outcomes were worse for patients with BSFs, with significantly increased mortality, acute CDI, fewer ventilator-free days, and longer PICU and hospital LOS. As BSFs themselves should not significantly increase mortality, the presence of BSFs is likely indicative of a greater force of impact required to generate the fracture. Indeed, vascular trauma and meningitis do not adequately account for the increased mortality rate among our patients with BSFs. Moreover, SAH was associated with BSFs in this study, but SAH was not independently associated with mortality in this same cohort of sTBI patients.29 Mortality rates in children with sTBI range from 5% to 51%,5,6 although most recent studies report mortality rates in the range of 10% to 30%.10,25Y29,46Y48 Survival carries a significant risk of severe disability. Many studies have addressed premorbid, clinical, radiologic, and biochemical variables that may allow clinicians to predict individual patient outcome, although none is reliable in isolation. Mortality is uncommon (È1.6Y14%) in an inclusive trauma population of children with BSFs,19,49 although, in our cohort of sTBI patients with BSFs, the mortality rate was 45%. Our analysis indicated that BSFs could be added to other reported mortality signals, including acute CDI,26,27,29 absent pupillary responses,26,28,29 low motor GCS scores,28,29,50 hypotension,28,29 sodium dysregulation,26 and the presence of ischemia29 or cerebral edema27 on admission head CT. There were several limitations to our study. First, ours was a retrospective single-center study. Nonetheless, our data should be relevant for other Level 1 trauma centers in developed countries, as we used strict inclusion criteria for sTBI and standard anatomic definitions for a BSF. Second, the retrospective nature of this study raises the possibility of missed data points; however, we used our provincially mandated, qualitycontrolled trauma registry to capture most data. Third, vascular injuries in our study may be underreported because most of our data were acquired before pediatric trauma centers adopted the Modified Memphis Criteria for screening of blunt cerebrovascular injury. Fourth, our database only captures
* 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
1159
J Trauma Acute Care Surg Volume 78, Number 6
Alhelali et al.
short-term outcomes. Thus, future studies should aim to determine associations between BSFs and functional long-term outcomes in sTBI survivors (i.e., irreversible cranial nerve injury, hearing impairment). In summary, we report a BSF prevalence of 26% in pediatric sTBI patients. Importantly, BSFs were strongly associated with mortality. Pediatric sTBI patients with BSFs were also more likely to develop acute CDI and CSF leaks, as well as fewer ventilator-free days. Survivors of sTBI who had BSFs experienced longer PICU and hospital LOS. Overall, our data suggest that poor short-term outcomes were associated with BSFs. BSFs likely represent the significant linear force of impact and a significant mortality signal in pediatric sTBI. AUTHORSHIP All authors made substantial contributions to the conception, study design, content, and revision of the manuscript. I.A., I.M.A., and H.D. collected and entered patient data. I.A. and J.F. performed the literature search. T.C.S. performed the epidemiologic analyses. I.A. drafted the initial manuscript. A.R. and D.D.F. supervised this project, provided the medical interpretation of the data and results, and critically reviewed and revised the manuscript. All authors approved the manuscript as it has been submitted.
ACKNOWLEDGMENT The authors thank ‘‘LW Stitt Statistical Services’’ for independent methodological and statistical review.
DISCLOSURE D.D.F. is supported by the Children’s Health Foundation (www.childhealth.ca) and the Centre for Critical Illness Research in London, Ontario, Canada. The other authors declare no conflicts of interest.
REFERENCES 1. Shi J, Xiang H, Wheeler K, Smith GA, Stallones L, Groner J, Wang Z. Costs, mortality likelihood and outcomes of hospitalized US children with traumatic brain injuries. Brain Inj. 2009;23(7):602Y611. 2. Chung CY, Chen CL, Cheng PT, See LC, Tang SF, Wong AM. Critical score of Glasgow Coma Scale for pediatric traumatic brain injury. Pediatr Neurol. 2006;34(5):379Y387. 3. Jimenez N, Ebel BE, Wang J, Koepsell TD, Jaffe KM, Dorsch A, Durbin D, Vavilala MS, Temkin N, Rivara FP. Disparities in disability after traumatic brain injury among Hispanic children and adolescents. Pediatrics. 2013; 131(6):e1850Ye1856. 4. Morrison G, Fraser DD, Cepinskas G. Mechanisms and consequences of acquired brain injury during development. Pathophysiology. 2013; 20(1):49Y57. 5. Bruce DA, Schut L, Bruno LA, Wood JH, Sutton LN. Outcome following severe head injuries in children. J Neurosurg. 1978;48(5):679Y688. 6. Kraus JF, Fife D, Conroy C. Pediatric brain injuries: the nature, clinical course, and early outcomes in a defined United States’ population. Pediatrics. 1987;79(4):501Y507. 7. Tomberg T, Rink U, Pikkoja E, Tikk A. Computerized tomography and prognosis in paediatric head injury. Acta Neurochir (Wien). 1996;138(5): 543Y548. 8. Claret Teruel G, Palomeque Rico A, Cambra Lasaosa FJ, Catala` Temprano A, Noguera Julian A, Costa Clara` JM. Severe head injury among children: computed tomography evaluation as a prognostic factor. J Pediatr Surg. 2007;42(11):1903Y1906. 9. Hirsch W, Schobess A, Eichler G, Zumkeller W, Teichler H, Schluter A. Severe head trauma in children: cranial computer tomography and clinical consequences. Paediatr Anaesth. 2002;12(4):337Y344.
1160
10. Feickert HJ, Drommer S, Heyer R. Severe head injury in children: impact of risk factors on outcome. J Trauma. 1999;47(1):33Y38. 11. Ong L, Selladurai BM, Dhillon MK, Atan M, Lye MS. The prognostic value of the Glasgow Coma Scale, hypoxia and computerised tomography in outcome prediction of pediatric head injury. Pediatr Neurosurg. 1996;24(6):285Y291. 12. Levin HS, Aldrich EF, Saydjari C, Eisenberg HM, Foulkes MA, Bellefleur M, Luerssen TG, Jane JA, Marmarou A, Marshall LF, et al. Severe head injury in children: experience of the Traumatic Coma Data Bank. Neurosurgery. 1992;31(3):435Y443. 13. 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. 2014;13(3):307Y314. 14. Einhorn A, Mizrahi EM. Basilar skull fractures in children. The incidence of CNS infection and the use of antibiotics. Am J Dis Child. 1978;132(11): 1121Y1124. 15. Dahiya R, Keller JD, Litofsky NS, Bankey PE, Bonassar LJ, Megerian CA. Temporal bone fractures: otic capsule sparing versus otic capsule violating clinical and radiographic considerations. J Trauma. 1999;47(6): 1079Y1083. 16. Raskind R, Doria A. Cerebrospinal fluid rhinorrhea and otorrhea of traumatic origin. Int Surg. 1966;46(3):223Y227. 17. Johnson K, Fischer T, Chapman S, Wilson B. Accidental head injuries in children under 5 years of age. Clin Radiol. 2005;60(4):464Y468. 18. Smirniotopoulos J, Mirvis S, Lefkowitz D. Imaging of craniocerebral trauma. In: Mirvis S, Shanmuganathan K, eds. Imaging in Trauma and Critical Care. 2nd ed. Philadelphia, PA: Saunders; 2003:37Y38. 19. Liu-Shindo M, Hawkins DB. Basilar skull fractures in children. Int J Pediatr Otorhinolaryngol. 1989;17(2):109Y117. 20. Kadish HA, Schunk JE. Pediatric basilar skull fracture: do children with normal neurologic findings and no intracranial injury require hospitalization? Ann Emerg Med. 1995;26(1):37Y41. 21. Kitchens JL, Groff DB, Nagaraj HS, Fallat ME. Basilar skull fractures in childhood with cranial nerve involvement. J Pediatr Surg. 1991;26(8): 992Y994. 22. Fontela PS, Tampieri D, Atkinson JD, Daniel SJ, Teitelbaum J, Shemie SD. Posttraumatic pseudoaneurysm of the intracavernous internal carotid artery presenting with massive epistaxis. Pediatr Crit Care Med. 2006;7(3): 260Y262. 23. Probst R, Fiebach A, Moser A. Fronto-basal fractures in the child [In German]. Laryngorhinootologie. 1990;69(3):150Y154. 24. Ratilal BO, Costa J, Sampaio C. Antibiotic prophylaxis for preventing meningitis in patients with basilar skull fractures. Cochrane Database Syst Rev. 2011;(8):CD004884. 25. Alharfi IM, Charyk Stewart T, Al Helali I, Daoud H, Fraser DD. Infection rates, fevers, and associated factors in pediatric severe traumatic brain injury. J Neurotrauma. 2014;31(5):452Y458. 26. Alharfi IM, Stewart TC, Kelly SH, Morrison GC, Fraser DD. Hypernatremia is associated with increased risk of mortality in pediatric severe traumatic brain injury. J Neurotrauma. 2013;30(5):361Y366. 27. Alharfi IM, Stewart TC, Foster J, Morrison GC, Fraser DD. Central diabetes insipidus in pediatric severe traumatic brain injury. Pediatr Crit Care Med. 2013;14(2):203Y209. 28. Stewart TC, Alharfi IM, Fraser DD. The role of serious concomitant injuries in the treatment and outcome of pediatric severe traumatic brain injury. J Trauma Acute Care Surg. 2013;75(5):836Y842. 29. Hochstadter E, Stewart TC, Alharfi IM, Ranger A, Fraser DD. Subarachnoid hemorrhage prevalence and its association with short-term outcome in pediatric severe traumatic brain injury. Neurocrit Care. 2014;21(3):505Y513. 30. Hutchison JS, Ward RE, Lacroix J, He´bert PC, Barnes MA, Bohn DJ, Dirks PB, Doucette S, Fergusson D, Gottesman R, et al., Hypothermia Pediatric Head Injury Trial Investigators and the Canadian Critical Care Trials Group. Hypothermia therapy after traumatic brain injury in children. N Engl J Med. 2008;358(23):2447Y2456. 31. Ngo QN, Ranger A, Singh RN, Kornecki A, Seabrook JA, Fraser DD. External ventricular drains in pediatric patients. Pediatr Crit Care Med. 2009;10(3):346Y351.
* 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Trauma Acute Care Surg Volume 78, Number 6
Alhelali et al.
32. Chan DT, Poon WS, Ip CP, Chiu PW, Goh KY. How useful is glucose detection in diagnosing cerebrospinal fluid leak? The rational use of CT and Beta-2 transferrin assay in detection of cerebrospinal fluid fistula. Asian J Surg. 2004;27(1):39Y42. 33. Available at: http://www.cdc.gov/hai/vap/vap.html. Accessed April 22, 2015. 34. Available at: http://www.cdc.gov/HAI/ca_uti/uti.html. Accessed April 22, 2015. 35. Available at: http://www.cdc.gov/hai/bsi/bsi.html. Accessed April 22, 2015. 36. Hosmer D, Lemeshow S. Applied Logistic Regression. New York, NY: Wiley; 1989;xiii:307. 37. Medana IM, Esiri MM. Axonal damage: a key predictor of outcome in human CNS diseases. Brain. 2003;126(Pt 3):515Y530. 38. Gurdjian ES, Roberts VL, Thomas LM. Tolerance curves of acceleration and intracranial pressure and protective index in experimental head injury. J Trauma. 1966;6(5):600Y604. 39. Perheentupa U, Kinnunen I, Gre´nman R, Aitasalo K, Ma¨kitie AA. Management and outcome of pediatric skull base fractures. Int J Pediatr Otorhinolaryngol. 2010;74(11):1245Y1250. 40. Michaud LJ, Rivara FP, Grady MS, Reay DT. Predictors of survival and severity of disability after severe brain injury in children. Neurosurgery. 1992;31(2):254Y264. 41. Rivara FP, Koepsell TD, Wang J, Temkin N, Dorsch A, Vavilala MS, Durbin D, Jaffe KM. Disability 3, 12, and 24 months after traumatic brain injury among children and adolescents. Pediatrics. 2011;128(5):e1129Ye1138. 42. White JR, Farukhi Z, Bull C, Christensen J, Gordon T, Paidas C, Nichols DG. Predictors of outcome in severely head-injured children. Crit Care Med. 2001;29(3):534Y540.
43. Lew SM, Frumiento C, Wald SL. Pediatric blunt carotid injury: a review of the National Pediatric Trauma Registry [Review]. Pediatr Neurosurg. 1999;30(5):239Y244. 44. McGuirt WF Jr, Stool SE. Cerebrospinal fluid fistula: the identification and management in pediatric temporal bone fractures. Laryngoscope. 1995;105(4 Pt 1):359Y364. 45. Yilmazlar S, Arslan E, Kocaeli H, Dogan S, Aksoy K, Korfali E, Doygun M. Cerebrospinal fluid leakage complicating skull base fractures: analysis of 81 cases. Neurosurg Rev. 2006;29(1):64Y71. 46. Hutchison J, Ward R, Lacroix J, He´bert P, Skippen P, Barnes M, Meyer P, Morris K, Kirpalani H, Singh R, et al., HyP-HIT Investigators, Canadian Critical Care Trials Group. Hypothermia pediatric head injury trial: the value of a pretrial clinical evaluation phase. Dev Neurosci. 2006; 28(4Y5):291Y301. 47. Campbell CG, Kuehn SM, Richards PM, Ventureyra E, Hutchison JS. Medical and cognitive outcome in children with traumatic brain injury. Can J Neurol Sci. 2004;31(2):213Y219. 48. Adelson PD, Wisniewski SR, Beca J, Brown SD, Bell M, Muizelaar JP, Okada P, Beers SR, Balasubramani GK, Hirtz D, Paediatric Traumatic Brain Injury Consortium. Comparison of hypothermia and normothermia after severe traumatic brain injury in children (Cool Kids): a phase 3, randomised controlled trial. Lancet Neurol. 2013;12(6):546Y553. 49. Ort S, Beus K, Isaacson J. Pediatric temporal bone fractures in a rural population. Otolaryngol Head Neck Surg. 2004;131(4):433Y437. 50. Fortune PM, Shann F. The motor response to stimulation predicts outcome as well as the full Glasgow Coma Scale in children with severe head injury. Pediatr Crit Care Med. 2010;11(3):339Y342.
* 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
1161
Basal skull fractures are associated with mortality in pediatric severe traumatic brain injury Ibrahim Alhelali, MD, Tanya Charyk Stewart, MSc, Jennifer Foster, MD, Ibrahim M. Alharfi, MD, Adrianna Ranger, MD, Hani Daoud, MD, and Douglas D. Fraser, MD, PhD, London, Ontario, Canada
Basal skull fractures (BSFs) are caused by blunt force trauma, occurring in the temporal, occipital, sphenoid, and/or ethmoid bones. In pediatric severe traumatic brain injury (sTBI), there is a paucity of data on BSFs. Our goal was to investigate the BSF prevalence, anatomy, and association with short-term outcomes in pediatric sTBI. METHODS: We retrospectively reviewed all severely injured (Injury Severity Score Q12) pediatric patients (aged G18 years) admitted to our hospital after experiencing an sTBI (Glasgow Coma Scale score e8 and head Abbreviated Injury Scale score Q4). Neuroimaging for all sTBI patients was reviewed for skull fractures. Data were analyzed with both univariate and multivariate techniques. RESULTS: Of the 180 patients with sTBI, 47 had BSFs for a prevalence of 26% (69 BSFs in total; 16 sTBI patients had Q2 BSFs). The squamous temporal bone was fractured most frequently (n=30/47 sTBI patients with BSFs). Patients with BSFs were heavier and had more facial injuries than those without (p G 0.05) but were similar in all other admission demographics, injury profiles, and clinical characteristics. Cerebrospinal fluid leak was found in 32% (n = 15 of 47) of BSF patients (otorrhea, n = 12; rhinorrhea, n = 1; otorrhea/rhinorrhea, n = 2; p G 0.001). Mortality, acute central diabetes insipidus, and fewer ventilator-free days were associated with BSFs (p G 0.005), whereas in sTBI survivors, BSFs were associated with longer lengths of stay (p G 0.05). Multiple logistic regression showed that BSFs were positively associated with the presence of subarachnoid hemorrhage (odds ratio [OR], 4.00; p = 0.001), contusion (OR, 2.48; p = 0.029), herniation (OR, 3.40; p = 0.037), and cerebral edema (OR, 2.30; p = 0.047) but negatively associated with diffuse axonal injury (OR, 0.20; p = 0.003). BSFs and mortality were strongly associated (OR, 6.87; p = 0.019). CONCLUSION: BSFs occurred in 26% of pediatric sTBI patients. The temporal bone was fractured in two thirds of sTBI patients with BSFs, and one third was associated with cerebrospinal fluid leaks. BSFs represent a significant linear blunt force and are independent predictors of mortality. (J Trauma Acute Care Surg. 2015;78: 1155Y1161. Copyright * 2015 Wolters Kluwer Health, Inc. All rights reserved.) LEVEL OF EVIDENCE: Prognostic and epidemiologic study, level III. KEY WORDS: Traumatic brain injury; children; adolescents; skull fracture; mortality. BACKGROUND:
BACKGROUND Severe traumatic brain injury (sTBI) is a leading cause of death and disability among children.1Y4 Outcome prediction in the pediatric sTBI patients is difficult, and outcomes may vary significantly between centers.5,6 Abnormalities found on head imaging may aid sTBI prognostication,7Y9 particularly when combined with clinical data.6,10Y12 Skull fractures are relatively common findings on head imaging, reported in 43% to 62% of pediatric trauma patients.10,13 The frequency of basal skull fractures (BSFs) in the Submitted: December 10, 2014, Revised: February 26, 2015, Accepted: February 27, 2015. From the Departments of Paediatrics (I.A., J.F., I.M.A., H.D., D.D.F.) and Surgery (T.C.S., A.R.), Western University; Trauma Program (T.C.S.), London Health Sciences Center; and Children’s Health Research Institute (J.F., D.D.F.), London, Ontario, Canada; Department of Pediatric Critical Care (I.M.A.), King Fahad Medical City, Riyadh, Saudi Arabia; and Translational Research Centre (D.D.F.); Physiology and Pharmacology (D.D.F.) and Clinical Neurological Sciences (D.D.F.), Western University; and Centre for Critical Illness Research (D.D.F.), London, Ontario, Canada. Address for reprints: Douglas D. Fraser, MD, PhD, Room C2-C82, Children’s Hospital, London Health Sciences Centre, 800 Commissioners Rd. E., London, Ontario, Canada N6A 5W9; email: [email protected]. DOI: 10.1097/TA.0000000000000662
general pediatric TBI population has been reported to be 4% to 14%7,14Y17 and in one pediatric sTBI cases series to be 20%.10 This latter publication did not evaluate BSFs specifically in outcome analyses. BSFs generally occur when a linear skull fracture with a vertical or oblique orientation extends to the base of the skull.18 The skull base consists of the basilar portion of the temporal, occipital, sphenoid, and ethmoid bones. A common clinical sign of BSF is subcutaneous bleeding occurring over the mastoid process or orbit. Of particular concern, BSFs can be associated with cranial nerve injury, vascular disruption, hearing loss, and cerebrospinal fluid (CSF) leak.19Y22 Dural tears can result from BSFs23 and may be a potential entry point for bacterial invasion and subsequent meningitis, but antibiotic prophylaxis remains controversial.24 Based on the paucity of data surrounding BSF in the pediatric sTBI population, the aims of this study were to determine the BSF prevalence, anatomic location, and their relationship to short-term outcomes.
PATIENTS AND METHODS The Health Sciences Research Ethics Board at Western University approved this study. We screened all pediatric trauma
J Trauma Acute Care Surg Volume 78, Number 6
1155
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Trauma Acute Care Surg Volume 78, Number 6
Alhelali et al.
patients (91 month and G18 years) during a 12-year period (January 2000YDecember 2011) with an Injury Severity Score (ISS) of 12 or higher. Patients with an sTBI were identified by a pre-sedation Glasgow Coma Scale (GCS) score of 8 or less and a head Maximum Abbreviated Injury Scale (MAIS) score of 4 or higher.25Y29 All sTBI patients were admitted to the pediatric intensive care unit (PICU) located in the Children’s Hospital, London Health Sciences Centre (CH-LHSC), which is a regional pediatric Level I trauma center for Southwestern Ontario. CH-LHSC serves a geographic area of 19,000 km2 with a pediatric population of more than 500,000. Patients who were admitted more than 12 hours after their injury were excluded. Patients were identified from the prospectively collected written and electronic admission records of the PICU, collated with the data of the provincially mandated LHSC Trauma Registry.25Y28 The patient data collected included demographics, injury data, neuroimaging studies, and outcomes. Pre-sedation GCS was determined at either the scene or referring hospital or on arrival to LHSC trauma center. The pupillary response and hypotension (systolic blood pressure [SBP] G70 mm Hg for infants, SBP G70 mm Hg + (2 * age) for toddlers and children younger than 10 years, or SBP G90 mm Hg for children aged 10 years or older) were recorded on arrival to our trauma room. The reports of computed tomography (CT) imaging were collected. Specific abnormalities documented by fellowshiptrained neuroradiologists included skull fractures, cerebral edema, diffuse axonal injury (DAI), subarachnoid hemorrhage (SAH), subdural hemorrhage (SDH), epidural hemorrhage (EDH), intraventricular hemorrhage (IVH), brain herniation, midline shift, cerebral contusion, and ischemia. Indications for acute CT angiography included BSFs extending to the carotid canal, the jugular foramen, and/or the foramen magnum and suspected injury to the sagittal sinus and/ or brain ischemia in a large vessel distribution. Magnetic resonance (MR) angiography and/or venography was used in follow-up or in cases of evolving ischemic lesion. During the study period, patients with sTBI in our institution were placed on a temperature-modulating blanket in supine position, with the head elevated to 30 degrees. Normothermia was targeted in those sTBI patients not enrolled in a hypothermia trial30 with administration of antipyretics, passive cooling, and/or use of the temperature-modulating blanket. Antibiotics were not routinely administered, except for the inconsistent prophylactic use of cefazolin after intracranial pressure (ICP) monitor placement.31 Adequate analgesia and sedation were obtained with opioid and benzodiazepine infusions, respectively. Raised ICP was generally managed as per published previously.25Y29 If the ICP is more than 20 mm Hg to 25 mm Hg for longer than 5 minutes or for a rapidly raising ICP: 1) Drain CSF for 5 minutes if external ventricular drain (EVD) in situ;31 2) Mannitol 0.5 g/kg intravenously for 20 minutes every 6 hours as needed; 3) 3% NaCl given in boluses of 1 mL/kg to 2 mL/kg for 5 minutes every 12 hours as needed; 4) Hyperventilation to a PaCO2 35 mm Hg; 5) Barbiturate infusion titrated to ICP and cerebral perfusion pressure; and 6) Neurosurgical consult for potential decompressive craniectomy. Mannitol and 3% NaCl were held if the measured osmolality was greater than 320 mOsm/L and 360 mOsm/L, respectively. Low cerebral perfusion pressure secondary to arterial hypotension 1156
and without raised ICP was managed as follows: 0.9% NaCl (or colloid) 10 mL/kg intravenously for 5 to 30 minutes as needed, followed by administration of inotropes/vasopressors. Severe TBI outcomes included in-hospital mortality, ventilator-free days (unventilated days in the first 28 days of admission), PICU and hospital lengths of stay (LOS), and discharge destination. Discharge destinations included chronic rehabilitation hospital, acute care hospital, or home. Discharge to a chronic rehabilitation hospital was a less favorable outcome because it indicated that the patient was less functionally independent on discharge. CSF leak was determined by the presence of BSF on head CT and documentation of a clinical CSF leak, with or without glucose testing.32 Acute central diabetes insipidus (CDI) was defined as polyuria (urine output 9 4 mL/kg per hour for children G70 kg; 9300 mL/h for adult-size children Q70 kg) for at least two consecutive hours, hypernatremia (serum Na 9145 mmol/L), high serum osmolality (9300 mOsm/kg), and low urine osmolality (G300 mOsm/kg) at the time of diagnosis.27 Confirmation of CDI was based on symptom reversal with administration of 1desamino-8-D-arginine vasopressin (DDAVP) and/or vasopressin. Infections in sTBI patients were described previously.25 Culture-positive urinary tract infections (single organism; Q105 colony-forming units/mL) were determined within the first 48 hours of admission. Nosocomial infections recorded were identified by Centers for Disease Control criteria and included ventilator-associated pneumonia,33 catheter-associated urinary tract infection,34 and central line bloodstream infection.35 Meningitis and wound infections were confirmed by positive cultures of a single organism. Continuous variables were found to deviate from normality and, hence, medians with interquartile ranges (IQRs) were presented. Between-group comparisons were made using the Mann-Whitney U-test. For categoric variables, the frequencies and percentages were presented and between-group comparisons were made using the Pearson’s W2 or Fisher’s exact test where appropriate. PICU LOS, hospital LOS and postdischarge destination were assessed for survivors only. For all analyses, a value of p G 0.05 was considered statistically significant. Multivariate logistic regression modeling was performed with BSF as the outcome variable. Variables identified a priori to be assessed for inclusion in the model were age, gender, ISS, motor vehicle collision, abnormalities identified on admission head CT (DAI, cerebral edema, contusion, ischemia, herniation, midline shift, EDH, SDH, SAH, and IVH), and the presence of a serious facial injury (defined as MAIS 3Y6 in the face). Variables found to be significant in univariate analyses at the 0.25 level were then entered and allowed to be removed from the model at the 0.10 level in a backward elimination strategy. Sensitivity analyses were performed using different modeling methods (i.e., forward stepwise logistic regression).36 A second multivariate logistic regression analysis was undertaken to determine the association of patient’s demographics, physiologic variables, neuroimaging abnormalities, including BSF, and overall injury severity on mortality, while controlling for the possible confounding effects of these variables on the relationships. Possible confounders were identified a priori. Variables considered included age (assessed both as a continuous variable and dichotomized into adolescent [aged 13Y17 years] or * 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Trauma Acute Care Surg Volume 78, Number 6
Alhelali et al.
nonadolescent [aged G12 years]), gender, ISS, hypotension on arrival to the hospital, pre-sedation motor GCS, fixed pupillary response on hospital admission, CDI, and all abnormalities identified on head CT. A backward elimination procedure was used in which variables considered a priori to be clinically important (i.e., ISS, BSF) were forced into the model. Other variables found to be significant in univariate analyses at the 0.25 level were then entered and allowed to be removed from the model at the 0.10 level. Sensitivity analyses were performed using different modeling methods (i.e., forward stepwise logistic regression).36 For both models, the Hosmer-Lemeshow statistic was calculated to evaluate the model fit and the C statistic was calculated to evaluate the predictive accuracy of our logistic regression model. Correlations among independent variables were used to assess potential multicollinearity. Variables with strong collinearity were removed, and sensitivity analyses were performed. The final models were used to determine the estimated odds ratio (OR) for either BSF or mortality, respectively, for each variable adjusted for confounders. All analyses were performed using IBM SPSS Statistics Version 21 (IBM Corporation, Armonk, NY).
RESULTS Of 818 pediatric trauma patients screened, 180 cases were identified that met study inclusion criteria for sTBI (GCS score e8 and MAIS score Q 4). All sTBI patients were admitted to the PICU. Skull fractures, both vault and BSF, were found in 39% of sTBI patients (n = 71 of 180). A total of 69 BSFs were found in the 47 sTBI patients, with the squamous temporal bone fractured most frequently (n = 30 of 47 sTBI patients; Fig. 1). Sixteen sTBI patients had two or more fractures, with the most common fracture combination being in temporal and occipital bones (n = 7 patients). CT angiography was completed in 7 of the 47 patients with BSFs. Indications for CT angiography included fractures extending to the carotid canal (n = 5), the foramen magnum (n = 2), and the jugular foramen (n = 1). MR angiography (n = 2) and MR venography (n = 1) were used in follow-up.
No vascular lesions were identified in the patient population investigated. Patients in the BSF group had greater weight on admission and median facial MAIS scores than those without BSFs (p G 0.05), but the groups were otherwise similar with respect to admission demographics, injuries, and clinical characteristics (Table 1). Patients with BSFs also had a greater number of other admission head CT abnormalities, including SAH, cerebral edema, contusion, and herniation (p G 0.005; Table 2). In contrast, DAI was more common in patients without BSFs (p G 0.05; Table 2). There were significantly higher rates of mortality and acute CDI in patients with BSFs, as well as fewer ventilatorfree days (p e 0.003; Table 3). CSF leaks occurred only in sTBI patients with BSFs (n = 15 of 47) and consisted of otorrhea (n = 12), rhinorrhea (n = 1), or otorrhea/rhinorrhea (n = 2). Two sTBI patients developed meningitis, but only one patient had a BSF. Decompressive craniectomy was performed on 28% of sTBI patients with BSFs (n = 13 of 47) as compared with 9% of sTBI patients without BSFs (n = 12 of 133; p = 0.001). For sTBI survivors, those with BSFs had greater PICU and hospital LOS (p G 0.05; Table 3). CSF leaking resolved spontaneously in all surviving patients, albeit 45% of sTBI patients with BSFs had acute placement of EVDs with intermittent CSF drainage as part of our ICP control protocol (n = 21 of 47). A multivariate logistic regression modeling analysis was conducted to examine the association of patient and injury variables with BSFs (Table 4). Variables were analyzed for all two-way interactions, and as none were found to be significant and they were removed from the model. Our sensitivity analysis revealed two nearly identical models with both forward stepwise and backward elimination and with no multicollinearity of the variables evident in either model. The strength of association for all variables was also very similar in both models. The only difference between the models was that age was included in the backward elimination, as age was not statistically associated with sustaining a BSF (OR, 1.07; p = 0.07). For the model performance evaluation, a test of the full model against
Figure 1. A schematic diagram illustrating 69 BSF locations and frequencies in 47 pediatric sTBI patients (16 patients had Q2 BSFs). * 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
1157
J Trauma Acute Care Surg Volume 78, Number 6
Alhelali et al.
TABLE 1. Admission Demographics, Injuries, and Clinical Criteria for sTBI Patients With and Without BSF Variable
BSF (n = 47) No BSF (n = 133)
Age, y Weight, kg* Male, n (%) Etiology, n (%) MVC Fall Assault/abuse Other Injury Severity Score Injury profile MAIS head MAIS neck MAIS face* MAIS chest MAIS abdomen MAIS extremity/pelvis MAIS external Presedation GCS score Presedation motor GCS score Absent pupillary response, n (%) Hypotension, n (%)
15 (8) 60 (55)* 32 (68)
12 (12) 45 (48)* 81 (61)
TABLE 3. Short-term Outcomes in sTBI Patients With and Without BSF p
0.235 0.038* 0.381 0.471
35 (75) 5 (11) 4 (9) 3 (6) 34 (17)
100 (75) 12 (9) 18 (14) 3 (2) 35 (17)
0.657
5 (0) 2.5 (3) 2 (2)* 3 (1) 2.5 (1) 3 (1) 1 (0) 4 (4) 2 (3) 18 (38) 16 (34)
5 (0) 2 (0) 1 (1)* 4 (1) 2 (2) 3 (1) 1 (0) 5 (4) 2 (3) 36 (27) 27 (20)
0.554 0.352 0.018* 0.347 0.686 0.174 0.422 0.218 0.352 0.149 0.058
Continuous variables are presented as median (IQR) and categoric variables as n (%). MVC, motor vehicle collision.
a constant-only model was statistically significant, indicating that the predictor variables reliably distinguished between patients with or without BSFs (Omnibus Test of model coefficients, X62 = 43.59, p G 0.001). There was no evidence of a lack of fit (Hosmer and Lemeshow X72 = 7.05, p = 0.424), with an overall classification success of BSF at 80% for the final model. The C statistic was 0.809, indicating good predictive accuracy of our model. SAH (OR, 4.00; p = 0.001), contusion (OR, 2.48; p = 0.029), herniation (OR, 3.40; p = 0.037), and cerebral edema (OR, 2.30; p = 0.047) were found to be positively associated with sustaining a BSF. In contrast, DAI was negatively associated with BSF (OR, 0.20; p = 0.003).
TABLE 2. Admission Head CT Abnormalities in sTBI Patients With and Without BSF (n = 171*) Head CT Abnormalities SAH, n (%)† Cerebral edema, n (%)† Contusion, n (%)† Herniation, n (%)† DAI, n (%)† Midline shift, n (%) SDH, n (%) IVH, n (%) Ischemia, n (%) EDH, n (%)
BSF (n = 47)
No BSF (n = 124)
p
29 (62)† 31 (66)† 23 (49)† 10 (21)† 6 (13)† 12 (26) 25 (53) 8 (17) 7 (15) 6 (13)
42 (34)† 49 (40)† 32 (26)† 7 (6)† 37 (30)† 19 (15) 53 (43) 32 (26) 10 (8) 11 (8)
0.001† 0.002† 0.004† 0.004† 0.022† 0.122 0.221 0.226 0.250 0.391
*Only 171 admission head CTs were available for review.
1158
Outcomes
BSF (n = 47) No BSF (n = 133)
Mortality, n (%)* 21 (45)* Central diabetes insipidus, n (%)* 15 (32)* Cerebrospinal fluid leak, n (%)* 15 (32)* Ventilator-free days, n (%)* 14 (21)* Infection, n (%)† 8 (17) For survivors only* n = 26* Hospital LOS* 30.5 (26)* PICU LOS* 10.5 (8)* Discharge to, n (%) Home 17 (65) Acute care 1 (4) Rehabilitation 8 (31)
29 (22)* 17 (13)* 0 (0)* 23 (18)* 22 (17) n = 104* 18 (27)* 5 (8)*
p 0.003* 0.003* G0.001* G0.001* 0.940 0.046* 0.018* 0.421
68 (65) 12 (12) 24 (23)
†Two patients had meningitis, but only one patient had a BSF of the temporal bone without CSF leak. Continuous variables are presented as median (IQR) and categoric variables as n (%).
A second multivariate logistic regression modeling analysis was conducted to examine the association of patient and injury variables, including the effect of sustaining a BSF, on mortality (Table 5). Variables were analyzed for two-way interactions, and as none were found to be significant, they were removed from the model. Acute CDI was correlated with GCS motor, so acute CDI was removed. There was no other multicollinearity between the other variables in the model. We also evaluated cranial vault fractures; however, they were correlated with BSFs. Because cranial vault fractures had the weaker strength of association with mortality and a higher p value, they were removed. Sensitivity analysis confirmed the effect of the different types of skull fractures on mortality. Logistic regression models that included (i) cranial vault fractures, (ii) BSFs, or (iii) all skull fractures showed that only the BSFs were statistically associated with mortality and, thus, only BSFs were included in the final model. The unadjusted OR of BSF on death was 2.90 (95% confidence interval [95% CI], 1.43Y5.88; p = 0.003). A test of the full model against a constant-only model was statistically significant, indicating that the set of predictor variables reliably distinguished between survivors and nonsurvivors (Omnibus Test of model coefficients, X72 = 135.002, p G 0.001). The Hosmer and
TABLE 4. Multivariate Logistic Regression Model With BSF as the Outcome Variable Variable SAH* DAI* Contusion* Herniation* Cerebral edema* Hypotension
A
SE
OR
95% CI
p
1.39* j1.64* 0.91* 1.22* 0.83* 0.86
0.42* 0.55* 0.42* 0.59* 0.42* 0.45
4.00* 0.20* 2.48* 3.40* 2.30* 2.36
1.75Y9.15* 0.07Y0.57* 1.10Y5.60* 1.08Y10.73* 1.01Y5.21* 0.97Y5.69
0.001* 0.003* 0.029* 0.037* 0.047* 0.057
Constant A = j2.43; SE = 0.42.
* 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Trauma Acute Care Surg Volume 78, Number 6
Alhelali et al.
TABLE 5. Multivariate Logistic Regression Model With Death as the Outcome Variable Variable
A
SE
OR
95% CI
p
Absent pupillary response* 3.42* 0.72* 30.47* 7.37Y126.00* G0.001* GCS motor* j1.58* 0.48* 0.21* 0.08Y0.53* 0.001* BSF* 1.93* 0.82* 6.87* 1.38Y34.34* 0.019* Hypotension* 1.64* 0.77* 5.14* 1.13Y23.34* 0.034* Ischemia* 2.15* 1.09* 8.58* 1.02Y72.13* 0.048* Injury Severity Score 0.06 0.04 1.06 0.995Y1.14 0.069 Cerebral edema 1.27 0.71 3.56 0.89Y14.27 0.073 Constant A = j4.11; SE = 1.61.
Lemeshow test indicated a good fit of the model (X82 = 6.510; p = 0.590), with overall classification success of death at 90%. The C statistic was 0.967, indicating excellent predictive accuracy of our model. The Wald criterion demonstrated that mortality was independently associated with BSF (OR, 6.87; p = 0.019), along with other variables, including absent pupillary response (OR, 30.47; p G 0.001), GCS motor (OR, 0.21; p = 0.001), hypotension on arrival (OR, 5.14; p = 0.034), and ischemia on admission head imaging (OR, 8.58; p = 0.048).
DISCUSSION In this study, we report the prevalence of skull fractures in a cohort of well-described pediatric sTBI patients.25Y29 Skull fractures were found in 39% of sTBI patients, with BSFs observed in 26%. Two or more BSFs were found in 9% of sTBI patients. About two thirds of BSFs occurred in the temporal bone, and one third of BSFs were associated with CSF leaks. BSF was positively associated with admission head CT abnormalities including SAH, contusion, herniation, and cerebral edema but negatively associated with DAI. Mortality was strongly associated with BSFs. To our knowledge, this is the first study to fully characterize a population of pediatric sTBI patients with BSFs. The prevalence of BSFs in patients with sTBI in this cohort was 26%, consistent with a previous case series.10 The prevalence of BSFs in the general pediatric trauma population is much lower at 4% to 14%,7,14Y17 albeit difficult to accurately ascertain because clinical features are not reliable indicators of their presence and many patients with mild TBI will not undergo a CT scan. Linear blunt force trauma causes BSFs. Our data indicate that BSFs were associated with SAH, contusion, cerebral edema, and herniation. In contrast, DAI was negatively associated with BSF, a finding that is consistent with angular, or rotational, head injuries.37 The majority of BSFs in our study were in the temporal bone, an area of the skull considered least capable of absorbing a blow and, thus, the most susceptible to fracture.38 The distribution of fracture locations was similar to that seen in other reports.19,39 Similarly, the gender distribution (approximately two-thirds male) for the entire sTBI population and those with BSF in particular is consistent with the trauma literature.6,40Y42 Patients with BSFs had higher MAIS face,
suggesting that those trauma patients with serious facial injuries should be screened for BSFs. BSFs are associated with vascular disruption, CSF leak, and meningitis.19Y22 We did not identify any sTBI patients with large vessel occlusion/dissection, a finding consistent with the rarity of large vessel injury in pediatric BSFs; carotid injury associated with BSFs were estimated at 0.09%.43 CSF leaks were found in approximately one third of sTBI patients with BSFs, with the majority of cases otorrhea. In sTBI survivors, all CSF leaks resolved spontaneously, which is likely a factor of anatomy, as otorrhea secondary to temporal bone fracture tends to resolve spontaneously.44 Leak resolution also may have been hastened by our frequent use of acute EVD drainage.45 Meningitis developed in two sTBI patients, but only one of these patients had a BSF. Hence, our data were unable to establish any relationships between BSF, CSF leaks, and infection. Meningitis could occur secondary to bacterial seeding, dural tears, and/or from placement of an EVD.31 In our sTBI cohort, patients with BSFs had a higher occurrence of SAH, cerebral edema, and herniation, all indicating a higher severity of injury than in those sTBI patients without BSFs. Virtually all clinical short-term outcomes were worse for patients with BSFs, with significantly increased mortality, acute CDI, fewer ventilator-free days, and longer PICU and hospital LOS. As BSFs themselves should not significantly increase mortality, the presence of BSFs is likely indicative of a greater force of impact required to generate the fracture. Indeed, vascular trauma and meningitis do not adequately account for the increased mortality rate among our patients with BSFs. Moreover, SAH was associated with BSFs in this study, but SAH was not independently associated with mortality in this same cohort of sTBI patients.29 Mortality rates in children with sTBI range from 5% to 51%,5,6 although most recent studies report mortality rates in the range of 10% to 30%.10,25Y29,46Y48 Survival carries a significant risk of severe disability. Many studies have addressed premorbid, clinical, radiologic, and biochemical variables that may allow clinicians to predict individual patient outcome, although none is reliable in isolation. Mortality is uncommon (È1.6Y14%) in an inclusive trauma population of children with BSFs,19,49 although, in our cohort of sTBI patients with BSFs, the mortality rate was 45%. Our analysis indicated that BSFs could be added to other reported mortality signals, including acute CDI,26,27,29 absent pupillary responses,26,28,29 low motor GCS scores,28,29,50 hypotension,28,29 sodium dysregulation,26 and the presence of ischemia29 or cerebral edema27 on admission head CT. There were several limitations to our study. First, ours was a retrospective single-center study. Nonetheless, our data should be relevant for other Level 1 trauma centers in developed countries, as we used strict inclusion criteria for sTBI and standard anatomic definitions for a BSF. Second, the retrospective nature of this study raises the possibility of missed data points; however, we used our provincially mandated, qualitycontrolled trauma registry to capture most data. Third, vascular injuries in our study may be underreported because most of our data were acquired before pediatric trauma centers adopted the Modified Memphis Criteria for screening of blunt cerebrovascular injury. Fourth, our database only captures
* 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
1159
J Trauma Acute Care Surg Volume 78, Number 6
Alhelali et al.
short-term outcomes. Thus, future studies should aim to determine associations between BSFs and functional long-term outcomes in sTBI survivors (i.e., irreversible cranial nerve injury, hearing impairment). In summary, we report a BSF prevalence of 26% in pediatric sTBI patients. Importantly, BSFs were strongly associated with mortality. Pediatric sTBI patients with BSFs were also more likely to develop acute CDI and CSF leaks, as well as fewer ventilator-free days. Survivors of sTBI who had BSFs experienced longer PICU and hospital LOS. Overall, our data suggest that poor short-term outcomes were associated with BSFs. BSFs likely represent the significant linear force of impact and a significant mortality signal in pediatric sTBI. AUTHORSHIP All authors made substantial contributions to the conception, study design, content, and revision of the manuscript. I.A., I.M.A., and H.D. collected and entered patient data. I.A. and J.F. performed the literature search. T.C.S. performed the epidemiologic analyses. I.A. drafted the initial manuscript. A.R. and D.D.F. supervised this project, provided the medical interpretation of the data and results, and critically reviewed and revised the manuscript. All authors approved the manuscript as it has been submitted.
ACKNOWLEDGMENT The authors thank ‘‘LW Stitt Statistical Services’’ for independent methodological and statistical review.
DISCLOSURE D.D.F. is supported by the Children’s Health Foundation (www.childhealth.ca) and the Centre for Critical Illness Research in London, Ontario, Canada. The other authors declare no conflicts of interest.
REFERENCES 1. Shi J, Xiang H, Wheeler K, Smith GA, Stallones L, Groner J, Wang Z. Costs, mortality likelihood and outcomes of hospitalized US children with traumatic brain injuries. Brain Inj. 2009;23(7):602Y611. 2. Chung CY, Chen CL, Cheng PT, See LC, Tang SF, Wong AM. Critical score of Glasgow Coma Scale for pediatric traumatic brain injury. Pediatr Neurol. 2006;34(5):379Y387. 3. Jimenez N, Ebel BE, Wang J, Koepsell TD, Jaffe KM, Dorsch A, Durbin D, Vavilala MS, Temkin N, Rivara FP. Disparities in disability after traumatic brain injury among Hispanic children and adolescents. Pediatrics. 2013; 131(6):e1850Ye1856. 4. Morrison G, Fraser DD, Cepinskas G. Mechanisms and consequences of acquired brain injury during development. Pathophysiology. 2013; 20(1):49Y57. 5. Bruce DA, Schut L, Bruno LA, Wood JH, Sutton LN. Outcome following severe head injuries in children. J Neurosurg. 1978;48(5):679Y688. 6. Kraus JF, Fife D, Conroy C. Pediatric brain injuries: the nature, clinical course, and early outcomes in a defined United States’ population. Pediatrics. 1987;79(4):501Y507. 7. Tomberg T, Rink U, Pikkoja E, Tikk A. Computerized tomography and prognosis in paediatric head injury. Acta Neurochir (Wien). 1996;138(5): 543Y548. 8. Claret Teruel G, Palomeque Rico A, Cambra Lasaosa FJ, Catala` Temprano A, Noguera Julian A, Costa Clara` JM. Severe head injury among children: computed tomography evaluation as a prognostic factor. J Pediatr Surg. 2007;42(11):1903Y1906. 9. Hirsch W, Schobess A, Eichler G, Zumkeller W, Teichler H, Schluter A. Severe head trauma in children: cranial computer tomography and clinical consequences. Paediatr Anaesth. 2002;12(4):337Y344.
1160
10. Feickert HJ, Drommer S, Heyer R. Severe head injury in children: impact of risk factors on outcome. J Trauma. 1999;47(1):33Y38. 11. Ong L, Selladurai BM, Dhillon MK, Atan M, Lye MS. The prognostic value of the Glasgow Coma Scale, hypoxia and computerised tomography in outcome prediction of pediatric head injury. Pediatr Neurosurg. 1996;24(6):285Y291. 12. Levin HS, Aldrich EF, Saydjari C, Eisenberg HM, Foulkes MA, Bellefleur M, Luerssen TG, Jane JA, Marmarou A, Marshall LF, et al. Severe head injury in children: experience of the Traumatic Coma Data Bank. Neurosurgery. 1992;31(3):435Y443. 13. 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. 2014;13(3):307Y314. 14. Einhorn A, Mizrahi EM. Basilar skull fractures in children. The incidence of CNS infection and the use of antibiotics. Am J Dis Child. 1978;132(11): 1121Y1124. 15. Dahiya R, Keller JD, Litofsky NS, Bankey PE, Bonassar LJ, Megerian CA. Temporal bone fractures: otic capsule sparing versus otic capsule violating clinical and radiographic considerations. J Trauma. 1999;47(6): 1079Y1083. 16. Raskind R, Doria A. Cerebrospinal fluid rhinorrhea and otorrhea of traumatic origin. Int Surg. 1966;46(3):223Y227. 17. Johnson K, Fischer T, Chapman S, Wilson B. Accidental head injuries in children under 5 years of age. Clin Radiol. 2005;60(4):464Y468. 18. Smirniotopoulos J, Mirvis S, Lefkowitz D. Imaging of craniocerebral trauma. In: Mirvis S, Shanmuganathan K, eds. Imaging in Trauma and Critical Care. 2nd ed. Philadelphia, PA: Saunders; 2003:37Y38. 19. Liu-Shindo M, Hawkins DB. Basilar skull fractures in children. Int J Pediatr Otorhinolaryngol. 1989;17(2):109Y117. 20. Kadish HA, Schunk JE. Pediatric basilar skull fracture: do children with normal neurologic findings and no intracranial injury require hospitalization? Ann Emerg Med. 1995;26(1):37Y41. 21. Kitchens JL, Groff DB, Nagaraj HS, Fallat ME. Basilar skull fractures in childhood with cranial nerve involvement. J Pediatr Surg. 1991;26(8): 992Y994. 22. Fontela PS, Tampieri D, Atkinson JD, Daniel SJ, Teitelbaum J, Shemie SD. Posttraumatic pseudoaneurysm of the intracavernous internal carotid artery presenting with massive epistaxis. Pediatr Crit Care Med. 2006;7(3): 260Y262. 23. Probst R, Fiebach A, Moser A. Fronto-basal fractures in the child [In German]. Laryngorhinootologie. 1990;69(3):150Y154. 24. Ratilal BO, Costa J, Sampaio C. Antibiotic prophylaxis for preventing meningitis in patients with basilar skull fractures. Cochrane Database Syst Rev. 2011;(8):CD004884. 25. Alharfi IM, Charyk Stewart T, Al Helali I, Daoud H, Fraser DD. Infection rates, fevers, and associated factors in pediatric severe traumatic brain injury. J Neurotrauma. 2014;31(5):452Y458. 26. Alharfi IM, Stewart TC, Kelly SH, Morrison GC, Fraser DD. Hypernatremia is associated with increased risk of mortality in pediatric severe traumatic brain injury. J Neurotrauma. 2013;30(5):361Y366. 27. Alharfi IM, Stewart TC, Foster J, Morrison GC, Fraser DD. Central diabetes insipidus in pediatric severe traumatic brain injury. Pediatr Crit Care Med. 2013;14(2):203Y209. 28. Stewart TC, Alharfi IM, Fraser DD. The role of serious concomitant injuries in the treatment and outcome of pediatric severe traumatic brain injury. J Trauma Acute Care Surg. 2013;75(5):836Y842. 29. Hochstadter E, Stewart TC, Alharfi IM, Ranger A, Fraser DD. Subarachnoid hemorrhage prevalence and its association with short-term outcome in pediatric severe traumatic brain injury. Neurocrit Care. 2014;21(3):505Y513. 30. Hutchison JS, Ward RE, Lacroix J, He´bert PC, Barnes MA, Bohn DJ, Dirks PB, Doucette S, Fergusson D, Gottesman R, et al., Hypothermia Pediatric Head Injury Trial Investigators and the Canadian Critical Care Trials Group. Hypothermia therapy after traumatic brain injury in children. N Engl J Med. 2008;358(23):2447Y2456. 31. Ngo QN, Ranger A, Singh RN, Kornecki A, Seabrook JA, Fraser DD. External ventricular drains in pediatric patients. Pediatr Crit Care Med. 2009;10(3):346Y351.
* 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Trauma Acute Care Surg Volume 78, Number 6
Alhelali et al.
32. Chan DT, Poon WS, Ip CP, Chiu PW, Goh KY. How useful is glucose detection in diagnosing cerebrospinal fluid leak? The rational use of CT and Beta-2 transferrin assay in detection of cerebrospinal fluid fistula. Asian J Surg. 2004;27(1):39Y42. 33. Available at: http://www.cdc.gov/hai/vap/vap.html. Accessed April 22, 2015. 34. Available at: http://www.cdc.gov/HAI/ca_uti/uti.html. Accessed April 22, 2015. 35. Available at: http://www.cdc.gov/hai/bsi/bsi.html. Accessed April 22, 2015. 36. Hosmer D, Lemeshow S. Applied Logistic Regression. New York, NY: Wiley; 1989;xiii:307. 37. Medana IM, Esiri MM. Axonal damage: a key predictor of outcome in human CNS diseases. Brain. 2003;126(Pt 3):515Y530. 38. Gurdjian ES, Roberts VL, Thomas LM. Tolerance curves of acceleration and intracranial pressure and protective index in experimental head injury. J Trauma. 1966;6(5):600Y604. 39. Perheentupa U, Kinnunen I, Gre´nman R, Aitasalo K, Ma¨kitie AA. Management and outcome of pediatric skull base fractures. Int J Pediatr Otorhinolaryngol. 2010;74(11):1245Y1250. 40. Michaud LJ, Rivara FP, Grady MS, Reay DT. Predictors of survival and severity of disability after severe brain injury in children. Neurosurgery. 1992;31(2):254Y264. 41. Rivara FP, Koepsell TD, Wang J, Temkin N, Dorsch A, Vavilala MS, Durbin D, Jaffe KM. Disability 3, 12, and 24 months after traumatic brain injury among children and adolescents. Pediatrics. 2011;128(5):e1129Ye1138. 42. White JR, Farukhi Z, Bull C, Christensen J, Gordon T, Paidas C, Nichols DG. Predictors of outcome in severely head-injured children. Crit Care Med. 2001;29(3):534Y540.
43. Lew SM, Frumiento C, Wald SL. Pediatric blunt carotid injury: a review of the National Pediatric Trauma Registry [Review]. Pediatr Neurosurg. 1999;30(5):239Y244. 44. McGuirt WF Jr, Stool SE. Cerebrospinal fluid fistula: the identification and management in pediatric temporal bone fractures. Laryngoscope. 1995;105(4 Pt 1):359Y364. 45. Yilmazlar S, Arslan E, Kocaeli H, Dogan S, Aksoy K, Korfali E, Doygun M. Cerebrospinal fluid leakage complicating skull base fractures: analysis of 81 cases. Neurosurg Rev. 2006;29(1):64Y71. 46. Hutchison J, Ward R, Lacroix J, He´bert P, Skippen P, Barnes M, Meyer P, Morris K, Kirpalani H, Singh R, et al., HyP-HIT Investigators, Canadian Critical Care Trials Group. Hypothermia pediatric head injury trial: the value of a pretrial clinical evaluation phase. Dev Neurosci. 2006; 28(4Y5):291Y301. 47. Campbell CG, Kuehn SM, Richards PM, Ventureyra E, Hutchison JS. Medical and cognitive outcome in children with traumatic brain injury. Can J Neurol Sci. 2004;31(2):213Y219. 48. Adelson PD, Wisniewski SR, Beca J, Brown SD, Bell M, Muizelaar JP, Okada P, Beers SR, Balasubramani GK, Hirtz D, Paediatric Traumatic Brain Injury Consortium. Comparison of hypothermia and normothermia after severe traumatic brain injury in children (Cool Kids): a phase 3, randomised controlled trial. Lancet Neurol. 2013;12(6):546Y553. 49. Ort S, Beus K, Isaacson J. Pediatric temporal bone fractures in a rural population. Otolaryngol Head Neck Surg. 2004;131(4):433Y437. 50. Fortune PM, Shann F. The motor response to stimulation predicts outcome as well as the full Glasgow Coma Scale in children with severe head injury. Pediatr Crit Care Med. 2010;11(3):339Y342.
* 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
1161
