不同昏迷量表对急性卒中伴意识障碍患者预后的预测  

Predicting the prognosis in patients with acute stroke accompanied with conscious disturbance with different coma scales

作　　者：高岱全[1] 宿英英[1] 张运周[1] 王琳[1] 高冉[1] 赵经纬[1] 李霞[1] 

机构地区：[1]首都医科大学宣武医院神经内科,北京100053

出　　处：《中国脑血管病杂志》2009年第12期620-625,共6页Chinese Journal of Cerebrovascular Diseases

基　　金：首都医学发展科研基金资助项目(2007-1043)

摘　　要：目的对比研究格拉斯哥昏迷评分(GCS)、格拉斯哥-列日昏迷评分(GLS)、格拉斯哥匹兹堡昏迷评分(GCSP)、全面无反应性量表(FOUR)、反应水平量表(RLS)5个昏迷量表,在预测急性卒中伴意识障碍患者预后的价值。方法发病≤3d的卒中伴意识障碍患者120例,于发病第3、5、7天,分别采用不同昏迷量表对其进行评估,评估者间的一致性检验采用加权Kappa检验,量表内容一致性采用Cronbachα检验;以发病3个月改良Rankin量表评估患者的预后为标准,判断5个量表预测预后的准确性,预后预测的辨别力采用受试者工作曲线。结果①评估者间一致性方面FOUR量表最好(Kappa值:0.647),其他依次是GCSP(0.628)、RLS(0.613)、GCS(0.554)和GLS(0.497)量表。②GCS、GLS、GCSP、FOUR量表的内容一致性均较好(RLS为单一项目,不进行此项分析),Cronbachα值分别为0.826、0.845、0.913和0.853。③各昏迷量表均对不良预后具有较好的辨别预测能力,ROC曲线下面积依次为FOUR(0.854)、GCS(0.844)、GLS(0.838)、GCSP(0.793)和RLS(0.775)。④预测不良预后的最佳分界值为GCS≤6分、GCSP≤17分、GLS≤9分、FOUR评分≤8分、RLS≥6分。以昏迷量表分界值预测不良预后,FOUR、GCS、GLS、RLS、GCSP量表的准确率分别为82.3%、81.2%、79.4%、78.1%、76.5%。结论GCS和FOUR量表更适合用于对急性卒中伴意识障碍患者预后的评估,前者为神经科医护人员所熟悉,而后者更容易培训和掌握。Objective To comparatively study the prognostic values of 5 coma scales in predicting patients with acute stroke accompanied with conscious disturbance, including the Glasgow coma scale （GCS） , Glasgow-Liege coma scale （GLS） , Glasgow-Pittsburgh coma scale （GCSP）, Full Outline of Un- responsiveness scale （FOUR）, and reaction level scale （RLS）. Methods One hundred twenty patients with stroke accompanied with conscious disturbance ≤3 days after the onset of symptoms were assessed respectively at day 3, 5, and 7 using different coma scales. The consistency test between the assessors was assessed using the weighted kappa test, and the consistency of scale contents was assessed using the Cronbachg α test; Taking the Rankin scale as the criterion, the prognosis of the patients was assessed at 3 months after symptom onset. The accuracy of predicting prognosis with the five scales was estimated. The receiver operator characteristic curve （ROC） was used to predict the discrimination of prognosis. Results ①The FOUR scale was the best （ Kappa 0. 647 ） in the consistency between the assessors, followed by GCSP （0. 628）, RLS （0.613）, GCS （0.554）, and GLS （0. 497）. ②The consistency of GCS, GLS, GCSP and FOUR contents was better （as a single item RLS did not perform this analysis）, and the Cronbach α values were 0.826, 0. 845, 0. 913, and 0. 853, respectively. ③All the coma scales had better distinguish and predictive abilities for the poor prognosis. The areas under ROC curve were as follows： FOUR （0. 854）, GCS （0. 844）, GLS （0. 838）, GCSP （0.793）, and RLS （0.775）. ④The best cut-off values for predicting poor prognosis were GCS ≤6, GCSP ≤17, GLS ≤9, FOUR ≤8, and RLS ≥6. The poor prognosis was predicted with the cut-off values of coma scales, and the accuracy rates of FOUR, GCS, GLS, RLS, and GCSP were 82. 3%, 81.2%, 79.4% , 78.1%, and 76.5%, respectively. Conclusion GCS and FOUR are more suitable for patients with acute stroke accompani

关 键 词：卒中 意识障碍 预后 昏迷量表 

分 类 号：R743.3[医药卫生—神经病学与精神病学]