A Study On The Coma Scales And Electroencephalogram For The Prediction Of The Comatose Patient

Objective The purpose of this study is to evaluating the value of different coma scales and the electroencephalogram Young classification in predicting the mortality both in hospital and in three months which caused by acute coma.Methods In this study, 30 cases of coma patients were arranged in the intensive care unit of neurology department of Affiliated Hospital of Nantong University from August 2008 to December 2009. When they arrived, they would received an onset Electroencephalogram(EEG) examination, the results of EEG would be classified according to Young grading standards. Glasgow coma scale(GCS), Glasgow-Pittsburgh Coma Scale(GCS-P), Full Outline of Unresponsiveness score(FOUR) would be recorded according to the acute coma patients conditions at the same time. The mortality in-hospital and within 3-months , also the short-term prognosis would be evaluated by the improvement of the modified Rankin Scale(mRS).Results Among the recruited patients, we have 14 male patients, 16 female patients, aged from 19 to 89 (67.17±20.269 years), Among the 30 patients , cerebral hemorrhage was diagnosed in 12 cases (40%), cerebral infarction was 8 cases (26.7%), 6 cases of intracranial infection (20%), and 2 cases of status epilepticus (6.7%), subarachnoid hemorrhage was diagnosed (6.7%). The mortality both in hospital and in three months was 8 cases(8%) and 12 cases(40%) respectively. The survival patients lived with disabilities. Logistic regression analysis showed that GCS (b=-1.276,P = 0.036, OR, 0.279,95% CI: 0.085 ~ 0.9), GCS-P (b=-1.012,P = 0.039, OR, 0.846,95% CI: 0.722 ~ 0.992), FOUR (b=-0.431,P = 0.015 , OR, 0.65,95% CI: 0.46 ~ 0.919), were negatively related to the mortality in hospital. EEG of the Young classification (b=1.203,P = 0.005, OR = 3.3,95% CI: 1.45 ~ 7.645) were positively related to the mortality in hospital. The age (P = 0.978), gender ( P = 0.061), etiology (P = 0.082) had no statistical significance on the prognosis of coma patients. Cox hazards regression prompted GCS (P = 0.021, RR = 0.42, 95% CI: 0.201 ~ 0.877), GCS-P (P = 0.004 , RR = 0.858, 95% CI: 0.772 ~ 0.953), FOUR (P = 0.025, RR = 0.814, 95% CI: 0.68 ~ 0.974), EEG of the Young classification (P = 0.000, RR = 2.45, 95% CI :1.583-3 .803,) were related to the three-month short-term prognosis. The area under the ROC curve of the studies had shown that, GCS-P (AUC = 0.781) was better than the GCS (AUC = 0.758),FOUR scale (AUC = 0.778),EEG Young classification (AUC = 0.753), in which GCS = 6 points, GCSP = 17 points, FOUR = 6 points, EEG grading scale for theⅢgrades shown the highest sensitivity and specificity. Clinical score combined EEG classification could not improve prediction.Conclusions GCS , GCS-P , FOUR and EEG coma Young grading were relevant with the clinical prognosis. The lower cinical score or higher grade of EEG in patients accompanied with higher mortality rate both in hospital and within three months. Comprehensive GCS score, GCSP score, FOUR score and EEG classification, GCSP prediction score were better than the other scales. Clinical scale score combined EEG classification could not improve the value in predicting the mortality rate both in hospital and within three months.