Outcome Predictors in Pediatric Head Trauma: A Study of Clinicoradiological Factors.

INTRODUCTION: Traumatic injuries are the leading cause of death and a major cause of disability among children. About 70%-80% of the accidental deaths in pediatric age group result directly from central nervous system lesions.
METHODS: The purpose of our study was to study all the patients of ≤18 years of age with head or spinal injury admitted in neurointensive care unit at our center, an apex trauma center in a developing country, between June 2009 and September 2011. We retrospectively analyzed various factors including type of injury, mode of injury, admission Glasgow coma score (in case of head injury), and mortality rate. OBSERVATIONS: The study population consisted of 264 injured children. Mean age was 8.3 ± 5.6 years (range 5 months to 18 years). Forty percent of patients were within 1-5-year age group. Head injury accounted for 89% of cases and 11% of cases were spinal injury patients. Low-velocity trauma was the most common mode of injury, accounting for 74% of the cases. The percentage of patients with mild, moderate, and severe head injury were 38%, 15%, and 47%, respectively, in the head injury group. Mortality in head injury patients was 18% and in spinal injury patients was 9%. Operative intervention was done in 56% of patients. Predictors of mortality included severe head injury, hospital stay <7 days, pneumothorax, the presence of hypotension, and deranged coagulation parameters.
CONCLUSIONS: Head injury is much more common than spinal injury in pediatric patients and fall from height being the most common mode of injury. Severe head injury, hospital stay <7 days, pneumothorax, presence of hypotension, and deranged coagulation parameters are predictors of poor outcome.

INTRODUCTION

Trauma remains one of the most common causes of death in all age groups, but this is especially true in the pediatric population. Traumatic injuries are the leading cause of death and a major cause of disability among children.[12] Greater than 45% of deaths in children aged 1–14 years in the United States are secondary to trauma.[3] The most common cause of death in pediatric trauma in the United States has been shown to be traumatic brain injury.[4] About 70%–80% of the accidental deaths result directly from central nervous system lesions.[5] However, predicting, in the early phase, the long-term neurologic outcome in head-injured children is still a challenge.

Severe brain injuries are quite rare, representing only 10% of head trauma.[6] In the 1990s, Sharples et al.[7] found that 30% of the deaths could be avoided provided that adequate management could be initiated early. In contrast, prompt and accurate assessment of the severity of injury and early initiation of critical care is of crucial importance for preventing these deaths. Secondary brain lesions from systemic origin develop early and contribute to worsening of primary brain lesion and neuronal loss related to primary impact forces. Among them, hypoxia and hypotension are major sources of secondary brain lesions that worsen outcome. They are clearly accessible to critical care management and must be corrected as early as possible. The deleterious influence of uncontrolled secondary brain insults of systemic origin must also be considered when trying to define predictive factors of outcome. The purpose of our study is to describe a large population of cranial trauma children treated in an apex trauma center using a standardized inhospital approach and to determine the early predictors of mortality.

METHODS

A retrospective chart review was performed on all pediatric patients, less than 18 years of age, who had cranial injuries as evaluated in the emergency room of an apex trauma facility between June 2009 and September 2011. Patients were included if there was any evidence of cranial injury on the basis of imaging findings warranting admission in intensive care unit.

Data collected and recorded included the patient's age, gender, the mechanism of injury, systolic blood pressure, prothrombin time/international normalized ratio (INR), and presence of polytrauma. In addition, Glasgow Coma Scale (GCS) score, injury description, and inhospital mortality were recorded. The type of neurological injury was determined based on the results of initial computerized tomography scans. Hypotension was defined as a blood pressure that for >5 min was below the 5th percentile for age based on published definitions. Coagulopathy was defined as INR ≥1.5. The relationships among age, INR (coagulopathy), temperature, GCS, and inhospital mortality were explored using Chi-square test. Univariate followed by multivariate logistic regression analyses were performed to find the factors associated with inhospital mortality. Patients with cranial injury were analyzed to find factors predicting mortality. The results are reported as odds ratio (95% CI). P < 0.05 was considered statistically significant. Statistical analysis was carried out using Stata 11.0 (College Station, Texas, USA).

RESULTS

Demographics

The study population consisted of 264 injured children. Mean age was 8.3 ± 5.6 years (range 5 months to 18 years). Forty percent of patients were within 1–5-year age group [Figure 1]. More than two-thirds of patients were male (69%) and females constituted 31% of patients. Head injury accounted for 89% of cases and 11% of cases were spinal injury patients.

Figure 1

Pie chart showing age distribution of patients

Mechanism of injury

Unlike adult trauma patients, most patients (74%) suffered from low-velocity trauma, whereas 26% of patients suffered from high-velocity trauma. Falls represented the predominant mechanisms of injury in preschool children, whereas motor vehicle accidents were the leading cause of trauma in older children. Ten percent of patients suffered from polytrauma.

Baseline status

Forty seven percent of patients suffered from severe head injury [Figure 2]. Six percent of patients presented with a GCS score of 3 [Figure 3]. Coagulopathy was present in 5.7% of patients at the time of admission. Total leukocyte count was abnormal at the time of admission in 56% of patients [Figure 4]. Early hypotension (hypotension within 24 h of admission) was seen in 24.2% of patients.

Figure 2

Pie chart showing distribution of patients according to severity of head injury

Figure 3

Pie chart showing distribution of severely head-injured patients according to Glasgow coma score

Figure 4

Pie chart of patients according to total leukocyte count

Inhospital course

Fifty one percent of patients required an emergency surgical procedure including evacuation of hematoma or decompressive craniectomy or both depending on imaging findings. Mortality rate was 18%. Mean Hospital stay was 6.6 ± 8.1 days (range – 1–65 days). Fifteen percent of patients developed ventilator-associated pneumonia.

Univariate analysis showed that mortality rate was significantly higher in severe head injury as compared to mild head injury (34.9% vs. 2%; P = 0.001) [Table 1]. Significantly higher mortality rates were also seen among patients who had coagulopathy (73.3% vs. 14.8%; P = 0.001), early hypotension (50% vs. 8%; P = 0.001), and culture-proven infection (36.8% vs. 9.6%; P = 0.001). Factors which were not associated significantly higher mortality rate include duration of hospital stay, requirement of operative intervention, mechanism of injury, occurrence of ventilator-associated pneumonia, presence of polytrauma, deranged total leukocyte count, and associated pneumothorax. On multivariate analysis, hospital stay <7 days, moderate and severe head injury, coagulopathy, pneumothorax, and presence of hypotension were found to be significantly associated with mortality.

Table 1

Analyses of various factors in head-injured patients

DISCUSSION

Despite the development of advanced trauma centers and other advances such as monitoring of cerebral perfusion pressure and jugular venous oxygenation, significant mortality persists in children with craniospinal trauma. This study was designed to investigate prognostic factors that may impact mortality in pediatric craniospinal trauma victims. We report a large cohort of children with craniospinal trauma seen over a 28-month period in a single apex trauma center. Most of the patients were transported directly from the scene of trauma to our center.

Most studies have suggested that among head-injured patients, children are more likely to survive and do well than are adults.[89] However, within the pediatric population, some investigators[910] but not all[11] claim that children <5 years of age have a poorer prognosis. However, those who did not find any correlation mentioned that the number of young children included in their studies was insufficient to draw meaningful conclusion. We found that age of patient has a significant bearing on the outcome, with mortality rate in children <1 year of age being 57% as compared to 12% in patients over 5 years of age although a number of patients below 1 year were shown in Figure 5. Rivara et al.[12] reported that boys have double the rate of brain injuries when compared with female counterparts. In our study, 69% of the patients were boys; it is consistent with this finding. Although gender differences were important in the rate of injury, they did not appear to adversely affect neurological outcome.

Figure 5

Bar graph showing mortality rates in different age groups

Fall from height is the most common cause of injury in our study. Motor vehicle accidents accounted for nearly one-fourth of cases. In 2% of patients, assault was the cause of trauma. This is in contrast to the other reports, in which high-velocity trauma is the most common mode.[1314] The mechanism of injury did not appear to influence neurological outcome. Patients with operable mass lesions did not demonstrate either a better or a worse neurological outcome. This is in concordance with findings in other series.[14]

Many investigators claim that the initial GCS is an excellent predictor of mortality,[1516] whereas others maintain that early prediction of outcome is difficult and recommend applying the GCS 6 h after admission.[2] However, we observed that the classification of head injury into mild, moderate, and severe on the basis of initial GCS is a good prognostic factor for predicting chances of mortality. However, within the severe head injury group patients, further classification into high (GCS 6–8), mid (GCS 4–5), and low (GCS 3) was not very reliable in predicting mortality. Others have also observed that although there are clear differences in outcome when comparing low (3–4) to higher (7–8) GCSs, the middle range (5–7) serves as a less accurate prognostic factor.[1417]

That hypotension is related to neurological outcome in head-injured patients, is a well-known concept. Many investigators have reported that hypotension, independent of injury severity or GCS, is an early predictor of death in the adult population.[17181920] Pigula et al. suggested that prehospital hypotension was significantly related to mortality in head-injured children.[21] They demonstrated that hypotension or hypoxia on admission quadruples the risk of mortality. They found 85% mortality rate in children with both hypotension and hypoxia and 61% mortality rate in children with hypotension alone, compared with 22% in normotensive and normoxic children. Our study also suggests that early hypotension significantly increases morbidity and predicts poor outcome. There are several proposed mechanisms regarding the etiology of secondary brain injury. The already injured brain has a significantly increased susceptibility to secondary injury. Head injuries disrupt the normal cerebral autoregulation causing central hypotension leading to cerebral hypoperfusion and ischemia.[14]

Patregnani et al., in a retrospective review in children with traumatic injuries treated at combat support hospitals, concluded that coagulopathy and shock on admission are common and independently associated with a high incidence of inhospital mortality.[22] Our study also concluded that coagulopathy is associated with significantly higher mortality.

Polytrauma was not found to be significantly associated with mortality in head-injured patients. However, pneumothorax was found to be significantly associated with mortality in head-injured patients in multivariate analysis.

CONCLUSIONS

In this large series of craniospinal trauma children treated in a single apex trauma center, we demonstrated that head injury is much more common than spinal injury and fall from height being the most common mode of injury. Outcome can be anticipated in the early phase with simple clinical data. Low GCS, initial hypotension, presence of coagulopathy, and young age determine outcome.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.