Outcome of severe traumatic brain injury at a critical care unit: a review of 87 patients

METHODS: Basic demographic, clinical, radiological and mechanism of injury data were recorded at admission and during ICU stay. The main outcome measure was survival or death. The outcome groups were compared for the injury severity, mean arterial pressure, serum glucose level, grade of diffuse axonal injury and the presence of mydriasis and anisocoria using the X2 test and the Fischers Exact test as appropriate.


Introduction
Traumatic brain injury (TBI) has been termed a "silent global epidemic", accounting forup to 30% of all trauma related deaths and is the leading cause of death in young males in developed countries (1,2). It is considered to be a dual insult comprising primary and secondary processes. The primary injury is the physical or anatomic damage caused at the time of injury producing vulnerable cells that are further compromised by secondary brain injury factors (3)(4)(5)(6). The factors commonly associated with secondary insults include hypotension, hypothermia, hypoxia, acidosis and raised intracranial pressure (1,3,5). When these are present, mortality rises to 35-36% (1,5) as documented in the Traumatic coma Data Bank study  and that by Jeremitsky et al. In an earlier study at the Kenyatta National Hospital (KNH), patients with a pulse rate of less than 60 per minute at the time of admission had a mortality of 100% while 69% of those with pulse rate greater than 120 per minute had poor outcome (7). In managing severe TBI therefore, avoiding secondary brain injury is key.
The accurate prediction of short term and long term outcomes is useful for communication with the patient's family and health care professionals. Of the commonly used outcome measures, the Glasgow outcome scale is the oldest. It consists of fi ve categories namely death, persistent negative state, severe disability, moderate disability and good recovery. The current study aimed to evaluate the outcome and predictors of outcome of severe traumatic brain injury at ICU of Kenyatta National Hospital.

Materials and methods
Patients: This prospective study included all the consecutive adult patients with severe traumatic brain injury admitted to the ICU of Kenyatta National Hospital between April and September 2005.They were examined by the fi rst author and scored according to Glasgow coma scale (GCS) on arrival. They then underwent computed tomography (CT scan) as soon as this was practical. The study excluded patients with GCS of greater than 8 as those were admitted to the general wards. Patients meeting the study criteria were resuscitated according to advanced trauma life support (ATLS) protocols with supervision of the attending surgeon. Further data was collected on the cause of TBI, age, gender, vital signs and pupillary responses. Anisocoria was defi ned as pupillary inequality of more than 1 mm and mydriasis as fi xed and dilated pupils of 4mm or more in diameter. Hypotension was defi ned as mean arterial pressure of less than 70mmHg. The biochemical parameters measured at admission and during ICU stay included arterial blood gases, acid-base state, serum glucose and electrolytes.
Outcome data included hospital complications, mortality, status at discharge from ICU and length of ICU/hospital stay recorded. The CT scans were reported by one radiologist in collaboration with the investigators and classifi ed according to Marshal's classifi cation. In this classifi cation, the status of the mesencephalic cisterns, the degree of midline shift in millimeters, and the presence or absence of one or more surgical masses are considered. Diffuse Injury I included all head Diffuse Injury I included all head Diffuse Injury I injuries where there was no visible pathology; Diffuse Injury II included all injuries in which the Diffuse Injury II included all injuries in which the Diffuse Injury II cisterns were present, the midline shift was less than 5 mm, and/or there was no high-or mixeddensity lesion of more than 25 cc; Diffuse Injury III included injuries with swelling where the cisterns were compressed or absent and the midline shift was 0 to 5 mm with no high-or mixed-density lesion of more than 25 cc; and Diffuse Injury IV included injuries with a midline shift of more than 5 mm and with no high-or mixed-density lesion of more than 25 cc The outcome status was based on the Glasgow Outcome Score which ranged from good recovery to good recovery to good recovery death in fi ve categories. Good recovery was defi ned as capacity to resume Good recovery was defi ned as capacity to resume Good recovery normal occupational activities, while with moderate disability the patient was independent moderate disability the patient was independent moderate disability and resumed only part of activities of daily living. In case of severe disability the patient was not evere disability the patient was not evere disability able to engage in most previous personal, social and work activities unless with some form of assistance, while in persistent vegetative state the ersistent vegetative state the ersistent vegetative state patient was not aware of surroundings or only partially responded to stimuli. Poor outcome in this study was defi ned as death, persistent vegetative state or severe disability.
Statistical analysis: Analysis of the data was carried out using SPSS version 12 program to derive descriptive statistics and frequency distributions. Categorical data were expressed in terms of proportions while continuous variables were expressed as means and standard deviations.
Subgroups (survival & death) were analyzed in a univariate way using student t-test, chi square test and Ficher exact tests as appropriate. Signifi cance was set at p less than 0.05.

Results
Patient and injury characteristics: Eighty seven patients were reviewed. This represented 14.3% of the 609 adult ICU admissions. There were 73 men (83.9%) and 14 women (16.1%) with a mean age of 34 ± 17 years. The demographic and clinical parameters at admission are shown in Table 1.
Majority of the patients were aged 25-39 years (50.6%), while 21% were 40-60 years. Five percent of the patients were aged over 60 years.
Trauma in majority of the patients was due to motor vehicle accidents (58.6%) followed by assaults in 32.2%. Three patients (3.4%) were received in coma having been found on the streets by policemen with no clear history about their injuries being availed (Table 2).
Extracranial pathology was present in 78 patients (91.6%). Maxillofacial injury was present in 53 patients (49.5%) and limb fractures were seen in 27 patients (25.2%). Twenty of the patients with limb fractures had operative fi xation and the rest were managed conservatively. Ten patients had a combination  Outcome evaluation: Forty fi ve patients died on the initial ICU admission while one with severe disability died on readmission to ICU from the wards (total mortality of 54%). On discharge from ICU, 19 (21.8%) patients had persistent vegetative state, 7(8.0%) severe disability, 7(8.0%) moderate disability and 9(10.3%) had good recovery.
A number of patients and clinical parameters were related to outcome. Poor outcome in the age group 25-39 was 52.3%, 40-60 years (23.1%), 15-24 years (16.9%) and for those over 60 years (7.7%). Both the patients with splenic ruptures were admitted in hypovolemic shock, operated on, and died. Glasgow coma scores of 3-5 correlated with a high mortality (76.7%) as shown in Table 3 and Figure 1. Seventy fi ve patients had blood sugar levels taken of which 50 (66.7%) patients had levels less than 10mmol/L while 25 (33.3%) patients had levels more than 10mmol/L. Eighty eight percent of the patients with blood glucose levels of more than 10mmol/L died (P=0.010) ( Table 3).
Hypotension was recorded among 16 patients (23%) on admission. The rest had normal pressure. Of the 71 patients who had head CT scanning done, 58 patients (81.7%) had diffuse axonal injury III/IV and mass lesions while 13 patients (18.3%) had diffuse axonal injury grade I/II. Mortality in axonal injury grade I/II and III/IV was 30% and 64% respectively. Anisocoria was present in 27 patients while rhinorrhoea was present in 15 patients and ortorrhoea in 9 patients. Five patients had both ortorrhoea and Rhinorrhoea and all died. Table 3 summarizes the group comparisons.

Discussion
The results of the current study indicate a very high morbidity associated with intensive care of severe TBI at KNH. This occurs in a cohort of very young patients. Extracranial injuries were recorded in 91.6% of victims with the face and limbs being the commonest. The mean age of 34+ 17 years with 70.1% of these patients aged 18 to 39 years has profound economic implications. Injury robs these individuals in active employment their daily earnings and invites relatives to pay high ICU and hospital bills. Severe TBI patients represented 14.3% of the ICU admissions during the six months study period and of this only 44% survived. Our result on patient age concur with previous studies (7)(8)(9). Jeremitsky and colleagues have reported a mean age of 34 years (5). The mortality rate of 54% in the current study is however much higher than in previous reports (1,5,10,11). The reported rate was 36% in the study by Jeremitsky et al (5) and 35% in the traumatic coma data bank study (1). The high mortality in this study could be explained by high frequency of multiple injuries and late presentation to hospital. Most patients in this study presented one to four hours after the injury with subsequent delays in starting defi nitive management. The rate and pattern of extracranial injuries (91.6%) in this study could likely be explained by the mechanism of injury. Motor vehicle accidents and mob justice accounted for 90.8% of the injuries.
In the current study a GCS of 3-5 was associated with poor outcome. The GCS scoring system, fi rst described by Teadale and Janet (12) to quantify the state of consciousness, correlates well with neurological outcome (12)(13)(14). Quigley et al. (15) retrospectively reviewed all the patients admitted to their trauma center over three years who had GCS of 3, 4 and 5, and found their overall survival to be low (12.5%). Combined with pupillary abnormalities and old age, survival was even lower (15). In the prospective study by Ritter et al (16), GCS score <8 was associated with death or vegetative state in 39% of patients with two reactive pupils but 85% of those with both pupil non reactive. In the current study the results on anisocoria and mydriasis were not signifi cant. Pupillary reactivity is correlated with cerebral blood fl ow. In the study by Ritter et al. patients with the bilateral mydriasis had cerebral blood fl ow There is a direct relationship between the four diagnostic categories of diffuse axonal injury on CT scan and the mortality rate (17). Patients suffering diffuse injury with no visible pathology (Diffuse Injury I) have the lowest mortality rate (10%), while the mortality rate in patients suffering diffuse injury with a midline shift (Diffuse Injury IV) is greater than 50% (17,18) . When used in conjunction with the traditional division of intracranial hemorrhages (extradural, subdural, or intracerebral), this categorization allows a much better assessment of the risk of intracranial hypertension and of a fatal or nonfatal outcome (18). Unfortunately intracranial pressure measurements are not routinely done at KNH and therefore it was not possible to assess the impact of the various intracranial hemorrhages with associated diffuse axonal injury on outcome. In the current study the mortality rate was 30% and 64% in patients with diffuse axonal injury I/II and III/ IV respectively.
The negative infl uence of secondary systematic insults (SSI) in severe TBI is now well documented (5-9). The SSI are now divided into subgroups of respiratory SSI (hypoxemia, hypercapnia, hypocapnia), Circulatory SSI (arterial hypotension and hypertension) and metabolic or electrolyte SSI (anemia, hyper or hypoglycemia hyponatraemia, hypo or hyperkalemia). In the current study, serum glucose levels higher than 10mmol/l correlated with poor outcome on analysis (P=0.010) which is consistent with other studies (5,6). Jeremitsky et al (5) found hyperglycemia in 40% of patients with severe TBI and this was associated with increased mortality and longer hospital stay. In neurotrauma high serum glucose is thought to exacerbate secondary brain injury. Probably hyperosmolality, lactic acid production, alterations in neuronal PH and increases in excitatory amino acids lead to neuronal injury.
Hypotension has been incriminated more than any other SSI in the worsening TBI. A multicenter study using information from traumatic coma Data Bank illustrated that blood pressure less than 80mmhg worsened outcome at 6 months based on Glasgow outcome scale (8). It has specifi cally been shown that hypotension occurring in the prehospital phase, at admission or in ICU signifi cantly increases mortality and poor outcome (10). Chestnut showed that the combination of hypotension and hypoxia resulted in a 75% mortality rate (19). In this study 16 patients (23%) had hypotension as evidenced by mean arterial pressure below 70mmhg. The hypotensive patients had longer ICU stay with a 75% mortality recorded in this group.
There is a great deal of research into the biochemical and pathophysiological processes at work locally within the injured brain, but the fact remains that current monitoring and intervention is based on systemic or global brain variables. Moreover, there may also be a burden of undetected or unreported systemic hypoxiaand or hypotension before admission to hospital. Although our data may only confi rm earlier research fi ndings and the opinion of most clinicians that low GCS and secondary brain injury are signs of poor prognosis, it is also inevitable that as monitoring equipment develops, ever greater quantities of this kind of data will be generated.

Conclusion
Severe traumatic brain injury is a frequent cause of ICU admission especially among young patients and is associated with high mortality and morbidity. The outcome of severe TBI is poor and this is infl uenced by demographic, clinical, and CT Scan fi ndings. Glasgow coma scale lower than 5, diffuse axonal injury IV and intracranial mass lesions and blood glucose higher than 10mmol/l correlated with poor outcome on univariate analysis in this study.