Coma-awake Prediction Model Of Coma Patients With Severe Traumatic Brain Injury

Background&Objectives:In recent years, with the development of the system of social first aid, treatment techniques, and intensive care techniques, the mortality of patients with severe traumatic brain injury significantly lower than before, but the rate of disability remains high. Traumatic brain injury is a serious and complex trauma which with the particularity of anatomy and physiology. Traumatic brain injury contains primary brain injury and secondary brain injury. Primary brain injury is a brain injury which was caused by the trauma of violence Immediately, such as:scalp injury, skull fracture, brain concussion, brain contusion and laceration, brain stem injury, hypothalamus injury, etc. Secondary brain injury formed in a period of time after brain damage, and its contain:brain edema, epidural hematoma, subdural hematoma, intracranial hematoma, etc. It is not only the primary injury of the central nervous system, and which will ensue a series of secondary injury to aggravate its condition. Each year in our country, there are about60million head injury occurred, of which about100,000people died, and resulting in direct and indirect economic losses of up to10billion. Tremendous pain and burden to families and society with the patients who was in a long-term coma or persistent vegetative state. The early prognosis of patients with severe traumatic brain injury is an important factor in making treatment decisions, resource allocation, classify patients, or communicating with family. For coma patients who after recovery, the probability of awake is the common concerns with the neurosurgeon and the family.Coma which caused by damage to the nerve center of the structure and function activities (awareness, feelings, and activities), is a serious disturbance of consciousness and a common clinical manifestations. Awake means that people can understand and judgment with the surrounding environment then make the right reaction. It must be based on the cognitive function of the brain hemispheres perfect interplay between reticular awakening mechanism. According to the presence or absence of consciousness content, the wakefulness can be divided into mind wakefulness and mindless wakefulness. The body through a variety of senses, receive appropriate stimulation to generate nerve impulses which though the non-specific ascending activation system and the cortex to generate consciousness, also called the cortex awakening. The nerve impulses which was generated by body could through the hypothalamus biological clock in the brainstem ascending reticular activating system (ARAS) and then generate awakening called the unconscious awakening, also called subcortical arousal. The damage of all above the anatomical structure and function, can cause varying degrees of disturbance of consciousness. Coma rarely lasts longer than2to4weeks and is followed either by VS or recovery of consciousness. VS sustain for a month or more, can be diagnosed as PVS. According to statistics, the head injury which was due to the car accident, fall, assault and the movement there are about seven million people per year worldwide, with which was in a long-term coma or persistent vegetative state add at least50,000-10million people in China each year, which have brought tremendous pain and burden to families and society. The early prognosis of patients with severe traumatic brain injury is an important factor in making treatment decisions, resource allocation, classify patients, or communicating with family.So far, a lot of research and statistical analysis has been dong in the various predictions about the prognosis of patients with head injury assessment factor. Some studies found that:the most important factor for the Glasgow Outcome Scale (GOS) is age; GCS score is the most meaningful to recover the function of the nervous system; pupil reaction is the best evaluation for the quality of life, somatosensory evoked potential (SSEP) has a very strong association with the awake of coma patients. All above the prediction evaluation and research methods are from the developed countries. But these studies may not necessarily in line with China’s national conditions according to the patient’s race, physical fitness, in which different regions, different medical environment. In china, most of such prediction based on univariate analysis, and lack of prediction and assessment by multifactorial analysis. Most of the results of such predictions are not be agreed byfthe clinical specialists. Today, the awake probability of coma patients with severe traumatic brain injury is often predicted by clinicians on their experience, and lack of quantitative indicators analysis.We studied the awake prediction of coma patients with severe traumatic brain injury. Multivariate logistic regression analysis was used to develop a awake prediction model which was based on a series of predictors for coma patients with severe traumatic brain injury. It can be able to assess the probability of the risks and benefits of the patients, which assisted clinical decision making, and has high accuracy. It has obviously advantages compare to the simple prediction made by the clinical experience. We retrospectively reviewed the complete clinical data of263coma patients with severe traumatic brain injury, and explored a variety of factors which associated with the coma patient to awake, and establish awake prediction model of coma patients.Objects&Methods:1. We retrospectively reviewed the complete clinical data of263coma patients with severe traumatic brain injury who admitted in the Department of Neurosurgery, Huadu hospital, the Southern Medical University, from May2010to July2012..207patients were male and59were female, the average age of the patient was38.65±17.11years.2. The case selection criteria:①patients aged>16years and has clear history of traumatic brain injury, patient was coma on admission and initial Glasgow coma score (GCS) is less than8;②patients at least have one head CT scan after injury.3. The case exclusion criteria:①ptients have intracranial lesions before injury, on admission for sedation or anesthesia.②the patients who had serious underlying diseases such as coronary heart disease, chronic bronchitis inflammation, diabetes, renal failure and so on.③the patients who gave up therapy.4. We collected and collated the clinical data of the263cases, including a total of7possible influence factors which included the age, the admission of the pupillary light reflex, the mGCS scores(movement GCS scores), the brainstem whether was pressed in the CT images, the opening time of eyes after treatment, the percentages of ischemic brain volume in CT images.5. Quantify the various factors and establish a database.6. Statistical analysis:Univariate analysis using R×C Chi-square test; multi-variabes analysis using Multinomial Logistic Regression analysis, and determine the influence factors of the prognosis of comatose patients after brain injury, and then establish the regression equation. The significance level is a=0.05.Results:1. After treatment, in263cases of coma patients with severe traumatic brain injury, there are152cases (57.8%) come to awake in one month,40cases (15.2%) come to awake within one month to three months, and71cases (27%) remain in coma.2. The univariate analysis 2.1The waking rate of male and female patients were compared by the Rx C Chi-square test, the result showed that the difference has no statistically significant.(X2=1.219, P=0.544)2.2The waking rate of the three age-groups coma patients with severe traumatic brain injury were compared by the Rx C Chi-square test, the result showed that the difference has statistically significant.(X2=37.959, P=0.000)2.3The waking rate of the two groups of pupillary light reflex were compared by the Rx C Chi-square test, the result showed that the difference has statistically significant.(X2=32.908, P=0.000)2.4The waking rate of the two groups of mGCS were compared by the Rx C Chi-square test, the result showed that the difference has statistically significant.(X2=18.921, P=0.000)2.5The waking rate of the two groups of the brainstem whether was pressed in the CT images were compared by the Rx C Chi-square test, the result showed that the difference has statistically significant,(X2=54.132, P=0.000)2.6The waking rate of the two groups of the opening time of eyes after treatment were compared by the Rx C Chi-square test, the result showed that the difference has statistically significant.(X2=139.454, P=0.000)2.7The waking rate of the two groups of the percentages of ischemic brain volume in CT images were compared by the Rx C Chi-square test, the result showed that the difference has statistically significant.(X2=99.185, P=0.000)3. The Multivariate analysis3.1We developed two prognostic models based on Independent predictors with Multivariate Logistic Regression analysis. Model A included age, the admission of the pupillary light reflex, the mGCS scores(movement GCS scores), the brainstem whether was pressed in the CT images; Model B included the opening time of eyes after treatment, the percentages of ischemic brain volume in CT images.3.2Clinical predictionThe data of clinical coma patients was substituted into the formula:,to predict the waking probability of coma patients.3.3The discrimination of the prediction modelLikelihood ratio chi-square test shows that Model A and Model B has statistically significant(p<0.05). Pearson Residual was used to assess the merits of the model fitting, all of the absolute value is not greater than2, especially in the Model B Pearson Residual absolute values are less than1, more closer the prediction result with the observed values. Model B fit better than A. Conclusions:1. In our study, we found that age, the admission of the pupillary light reflex, the mGCS scores(movement GCS scores), the brainstem whether was pressed in the CT images, the opening time of eyes after treatment, the percentages of ischemic brain volume in CT images, are independent factors to predict the awakening probability of coma patients with severe traumatic brain injury. Pearson Residual was used to assess the merits of the model fitting.. The prediction model fitting was good.2. The prediction models have well goodness of fit, and predict factors that used in the models are easily to obtain in primly hospitals, so it have a wide range of clinical application prospects. The prediction models will provide important reference value for the development of clinical decision of coma patients with severe traumatic brain injury in early stage.