Pattern and Outcome of Pediatric Traumatic Brain Injury: A Prospective Cohort Study at Hawassa University Comprehensive Specialized Hospital, Southern Ethiopia

Background Traumatic brain injury (TBI), a major public health problem, is the most common cause of death/disability in children. Glasgow coma scale is used to assess, decide treatment and follow up of TBI. TBI causes and outcome data are scarce from sub-Saharan Africa, non-existent from Ethiopia. We aimed to document pattern and predictors of childhood TBI outcome in a teaching hospital, Southern Ethiopia. Prospective cohort study was conducted from September 2017 to September 2018 among pediatrics TBI presented to Hawassa University Hospital. Data were collected by structured questionnaires and analyzed using SPSS version 20. Logistic regression was carried out and significant associations were declared at p-value of < 0.05.

documented patterns. Presence of increased ICP, hyperglycemia, severe TBI and CT findings of contusion, DAI/intracranial bleeding were predictors of poor outcome. Public awareness on road safety, childhood safety in preventing falls/animal injuries, closer follow-up of TBI cases for ICP and glycemic controls are recommended. Background Traumatic brain injury (TBI) is a brain injury that occurs following a blow to head, a fall, a bullet, a high speed crash or explosion injuries. TBI could be an open (penetrating) or closed type [1].
Childhood injury requires an immediate attention given its contribution to high childhood mortality and long term disabilities. Injuries contribute to 5.4% (265,000-348,000) of childhood deaths per year worldwide [2]. In 2015, injuries resulted in 25,000 deaths among Ethiopian children 0-14years of age [3]. TBI is a single, severe and most common form of injury in children [4].
Worldwide it is estimated that TBI affects 69million individuals each year. Low and middle income countries (LMICS) have three times high TBI burden than high income countries. Road traffic related head injuries were reported to be common in LMICS. Globally TBI is projected to be the third leading cause of death and injury by world health organization in 2020 [2,5,6]. Pediatric TBI is reported to be the common cause of injury related death, and it commonly follows road traffic accidents and falls [7,8]. Western studies documented TBI contributing for 8.3% of pediatric emergency department(ED) visits, and commonly with mild severity [9,10]. Reports from developing countries documented TBI to be a very common public health problem with varied burden figures, commonly milder severity and of TBI cases [6,21,22]. Mild TBI presents with concussion symptoms affecting physical, cognitive, and emotional (affective) domains. Various degrees of autonomic and neurologic dysfunction are seen in moderate and severe TBI cases in addition to mild TBI features [23,24]. Head computed tomography( CT) is recommended for children presenting with drowsiness or decreased mentation, any sign of basal skull fracture, focal neurologic deficit etc [25]. Several arguments are forwarded on abandonment of skull x-ray as an investigation means for TBI [26,27].
Acute management of TBI includes resuscitation and airway management, nutritional support, intubation for TBI with GCS< 8, follow up for increased intracranial pressure and other complications , and neurosurgical intervention [6]. Presence of cerebral edema, GCS< 8, hypoxemia, and hypernatremia were reported to be predictors of poor outcome among TBI cases [21,22,28].
Outcome of TBI varied in resource limited settings and neurocritical protocol for prehospital care was recommended [29]. Mortality rate ranged from 8% in western settings to 21.2% in developing regions [7,11,29]. Lack of prospective studies and injury data registries in most parts of Africa has made the assessment of TBI difficult [15]. No report is published in pediatric TBI in a prospective base from Ethiopia. Hence, we aimed at documenting the mechanisms of TBI, treatment modalities and outcome in a prospective study at tertiary hospital where neurosurgical interventions are possible.
Our study will fill the knowledge gap and assists policy makers, researchers and concerned stakeholders for planning detailed intervention.

Study Area
The study was conducted at Hawassa university comprehensive and specialized hospital (HUCSH), Hawassa, Ethiopia. Hawassa city is located 270 km South of Addis Ababa-Ethiopia's capital. HUCSH is the first and largest referral and teaching hospital in Southern Ethiopia. It serves a catchment population of over 18 million. The Pediatrics Department provides inpatient and outpatient services.
Neurosurgical interventions are provided by Neurosurgeons, radiologic imaging including head CT are read by radiologists. The hospital has an Intensive care unit (ICU) for care of critical ill patients.

Study design and period
A prospective cohort study was carried out among pediatric TBI cases, aged between 2months and 14years, and visited HUCSH from September 2017 to September 2018. Consecutively admitted TBI cases fulfilling the predefined criteria were included in the study after informed consent was obtained from family or guardians. Cases were excluded from the study when consent is not secured.

Sample size
The sample size was calculated using single proportion formula in Epi info 7.2 with assumptions of 95% confidence interval, power of 80%, percentage of outcome in exposed group -17%, odds ratio -4.2, and the final sample size was 331 [30].

Variables
The dependent variable was patient's outcome on discharge i.e. death, neurologic deficit or full recovery. Independent variables included socio-demographic data (age, sex, and place of residence), intent, mechanism and nature of injury, place of occurrence, severity of injury, associated extracranial injury, investigation and treatment type, and stay in the hospital.

Operational definitions
Traumatic brain injury (TBI) was defined as a brain injury that occurs following a blow to head, a fall, a bullet, a high speed crash or explosion injuries [1]. TBI severity was graded into mild, moderate and severe when GCS is 13-15, 9-13, and severe < 8 [6]. Outcome was assessed using the GOS (Glasgow outcome score) during hospital discharge by physicians in charge of patient care [21,22].
Hypotension and hypertension were considered when blood pressure for age and sex was below 50 th percentile and >95 th percentile [31]. Hypoglycemia and hyperglycemia were considered when admission random blood sugar is <70mg/dl and >200mg/dl [32, 33].

Data collection tool, procedure and data quality assurance
Data were collected at first presentation of the patient to pediatric emergency room/ department, and continued during admission and at the time of discharge by trained Bachelor of Science graduate nurses and intern doctors. Data were collected using structured questionnaire and information consisting of sociodemographic characteristics, characteristics of injury (i.e. intent, mechanism, nature, and place of injury) and patient's previous medical history, clinical work up and management, duration of admitted stay and discharge outcome were collected. All consecutive TBI cases seen at pediatric emergency department of HUCSH were included after consent was obtained from family or guardian. Daily collected data were checked by supervisor resident doctor for completeness and correction was made on spot and on daily basis.

Data analysis
Data were doubled entered into excel spread sheet and analyzed using SPSS version 20 software.
Descriptive statistics like percentage, mean and standard deviation were used for the presentation of demographic data and factors affecting outcome of pediatric head injury. Binary Logistic regression was used to assess the association between dependent and Independent variables. Variables with Pvalue < 0.05 were taken into multiple logistic regression models for controlling the possible effect of confounders and finally the variables which had independent association with outcome of TBI were identified on the basis of adjusted odds ratio(aOR), with 95% CI and P value <0.05.

Sociodemographic characteristics
There were 4258 pediatric emergency room(ER) visits and TBI contributed to 7.4% (317) of ER visits in one year study period. A total of 317 study subjects were included, 96% response rate. From 317 study subjects, there were 218(68.8%) males and 99 (31.2%) females with male to female ratio of 2.2:1. The study subjects aged from 7 months to 14 years; mean age of 7.66 + 3.82 years. School age groups, 5 -10 years and adolescents, 10-14 years contributed for 36.9% (117) and 36.3%(115) of TBI.
One hundred eighty three (57.7%) of study subjects came from southern nations, nationalities and peoples regional state (SNNPR) while the rest 134(42.3%) came from the neighboring Oromia and Somali Ethiopian regional states (See Table 1).
Boys were predominantly affected and they accounted for 65.8%, 68.3% and 71.3% of TBI in underfives, school age and adolescent age groups respectively.
Concerning timing of presentation, most of study subjects presented within 24 hours of injury, 258 (81.4%). Thirty one (9.8%) and 28(8.8%) of TBI cases presented between 24hours and 72hours, and after 72 hours of injury (See  Regarding duration of hospitalization, majority of TBI cases stayed from 4 to 7days, 34% (108) while 78(24.6%) were discharged with in 24hours of arrival. Prolonged hospitalization (≥ 1month) was seen in 2 patients with severe TBI and extremity fracture with extensive soft tissue injury. The average hospitalization days for cases who died was 4.5 days (median 2 days), and 5/ 10(50%) of deaths occurred within the first 3days of admission (See table 4).

Factors affecting outcome
On bivariate analysis factors significantly associated with outcome of pediatric TBI with 95% CI and p value < 0.05 were: comorbid illness, loss of consciousness and convulsion at presentation, increased ICP sign, severity of head injury, presence of hypotension, hyperglycemia on presentation, and head CT scan findings.

Discussion
In our study, boys and children above 5 years of age predominated. This finding is in agreement with Nigerian, South African and Tunisian studies. This could be related to boy's risk-prone behavior resulting in high energy transfer and their outdoor engagement [14,30,[34][35][36][37]. Concerning mechanism of injury, unintentional pedestrian RTA and falls were the commonest causes followed by intentional fighting/violence injuries. This is in line with most of the studies done in developing countries [2, 8, 18, 22,  In this study, subjects were evaluated with random blood sugar, skull X-ray and head CT scan at presentation. Most of the subjects had normal glucose level with hyperglycemia documented in 6.3% of TBI cases. Head CT scan was taken in 70.7% of study subjects and showed various types of skull bone fractures, brain contusion, intracranial bleeding and diffuse axonal injury. These findings are in agreement with studies done in India, Tunisia and Nigeria, and head CT scan requests conform to the recommended neurosurgical practice [22,25,30,37,39].
In our study, associated extracranial injuries were documented in 80% of the study subjects with soft tissue injury being the commonest followed by extremity bone fracture. Reports from Nigeria and Nepal documented similar findings [18,40] Severe TBI is associated with primary areflexia and secondary brain edema, which predicts the outcome of the patient[44]. Head CT scan findings of contusion, diffuse axonal injury and intracranial bleedings were associated with 2.45 times higher chances of death when compared with normal head CT cases. This is in agreement with Tunisian and Nigerian studies [22,30,39]. Diffuse axonal injury that occurs in TBI is reported to be secondary to axonal swelling, calcium mediated irreversible blockade of axonal transport, swollen endoplasmic reticulum etc. These will be evident on head CT as cerebral contusion, intracerebral hematoma and subarachnoid hemorrhages. It is reported these findings predict poor outcome among TBI cases with highest accuracy [45-47].
Our study documented 1.42 times higher chances of death and neurologic disability among TBI cases who presented with increased ICP signs when compared with those without increased ICP signs. This is in line with Tunisian and Argentinian studies [22,30,48]. Glial swelling with narrowed lumina of the microvasculature due to podocytic process swelling results in increased ICP. Diffuse cerebral ischemia also results in calcium hemostatic imbalance, activation of anaerobic metabolism. Increased ICP clinically presents as hypoxemia, seizure, mental level deterioration and neurologic deficits [49,50]

Conclusions
In our study, boys and children above 5years of age were highly affected by TBI. Pedestrian RTA, falls, fights, animal related injuries were the commonest mechanisms of injuries. Early presentation (<24hrs) and mild form of TBI were the commonest clinical presentations. Various forms of skull vault fractures: hemorrhage, contusion and axonal injuries were documented on head CT-scan. Majority of our study subjects were managed conservatively and recovered without neurologic deficits. Death was documented in 10(3.2%) of study subjects. Increased ICP and hyperglycemia at admission, severe TBI and head CT findings of contusion, diffuse axonal injury or intracranial bleeding were found to be predictors for TBI outcome. Public awareness on road safety needs to be strengthened. Parents or guardians need to be watchful in preventing fall and animal related pediatric injuries. Closer follow up of TBI cases for ICP and proper glycemic controls are recommended. Head CT-scan imaging should be considered in pediatric TBI.

Ethical approval and consent to participate
Ethical approval of the research protocol was obtained from institutional review board of Hawassa University. Written permission was secured from the hospital administration. Purpose of the study was explained to participants and consent was obtained from the parent or legal guardian of the child and assent was secured from adolescents. Confidentiality was assured by excluding patient names and by respecting the study subjects right not to participate or withdraw at any point from the study. Patient with no caretaker and who had severe head injury were excluded from study.

Consent for publication
Not applicable

Availability of data and material
The datasets analyzed during this study is available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.

Funding
Funding was received from Hawassa University, school of Graduate studies. The funding was used for the design and study data collection, analysis, interpretation of data and in writing the manuscript.
TB: wrote the draft proposal in consultation with HT, monitored data collection, entered data, analyzed, and wrote draft report. HT: conceived the idea and critically contributed to the design, analysis and interpretation of findings, wrote the current manuscript in its current form. Both authors have read and approved the manuscript for publication.   injury , CT % -Computer tomography, DAI^-diffuse axonal injury, ICH $ -intracranial hemorrhage Figures Figure 1 Mechanism of Pediatric traumatic brain injury at HUCSH from September 2017 to September