Reith Florence C M, Lingsma Hester F, Gabbe Belinda J, Lecky Fiona E, Roberts Ian, Maas Andrew I R
Department of Neurosurgery, Antwerp University Hospital and University of Antwerp, Edegem, Belgium.
Department of Public Health, Erasmus University Medical Center, Rotterdam, Netherlands.
Injury. 2017 Sep;48(9):1932-1943. doi: 10.1016/j.injury.2017.05.038. Epub 2017 Jun 1.
The Glasgow Coma Scale (GCS) is widely used in the assessment of clinical severity and prediction of outcome after traumatic brain injury (TBI). The sum score is frequently applied, but the differential influence of the components infrequently addressed. We aimed to investigate the contribution of the GCS components to the sum score, floor and ceiling effects of the components, and their prognostic effects.
Data on adult TBI patients were gathered from three data repositories: TARN (n=50,064), VSTR (n=14,062), and CRASH (n=9,941). Data on initial hospital GCS-assessment and discharge mortality were extracted. A descriptive analysis was performed to identify floor and ceiling effects. The relation between GCS and outcome was studied by comparing case fatality rates (CFR) between different component-profiles adding up to identical sum scores using Chi-tests, and by quantifying the prognostic value of each component and sum score with Nagelkerke's R derived from logistic regression analyses across TBI severities.
In the range 3-7, the sum score is primarily determined by the motor component, as the verbal and eye components show floor-effects at sum scores 7 and 8, respectively. In the range 8-12, the effect of the motor component attenuates and the verbal and eye components become more relevant. The motor, eye and verbal scores reach their ceiling-effects at sum 13, 14 and 15, respectively. Significant variations were exposed in CFR between different component-profiles despite identical sum scores, except in sum scores 6 and 7. Regression analysis showed that the motor score had highest R values in severe TBI patients, whereas the other components were more relevant at higher sum scores. The prognostic value of the three components combined was consistently higher than that of the sum score alone.
The GCS-components contribute differentially across the spectrum of consciousness to the sum score, each having floor and ceiling effects. The specific component-profile is related to outcome and the three components combined contain higher prognostic value than the sum score across different TBI severities. We, therefore, recommend a multidimensional use of the three-component GCS both in clinical practice, and in prognostic studies.
格拉斯哥昏迷量表（GCS）广泛用于评估创伤性脑损伤（TBI）后的临床严重程度和预后。总分常被应用，但各组成部分的差异影响却很少被提及。我们旨在研究GCS各组成部分对总分的贡献、各组成部分的地板效应和天花板效应及其预后效应。
收集来自三个数据库的成年TBI患者数据：TARN（n = 50,064）、VSTR（n = 14,062）和CRASH（n = 9,941）。提取初始医院GCS评估和出院死亡率的数据。进行描述性分析以确定地板效应和天花板效应。通过使用卡方检验比较不同组成部分组合且总分相同的情况下的病死率（CFR），并通过对不同TBI严重程度的逻辑回归分析得出的Nagelkerke's R来量化每个组成部分和总分的预后价值，研究GCS与预后的关系。
在3 - 7分范围内，总分主要由运动部分决定，因为言语和眼睛部分分别在总分7分和8分时显示出地板效应。在8 - 12分范围内，运动部分的影响减弱，言语和眼睛部分变得更重要。运动、眼睛和言语评分分别在总分13、14和15分时达到天花板效应。尽管总分相同，但不同组成部分组合之间的CFR存在显著差异，总分6分和7分的情况除外。回归分析表明，运动评分在重度TBI患者中具有最高的R值，而其他组成部分在总分较高时更重要。三个组成部分组合的预后价值始终高于单独的总分。
GCS各组成部分在意识范围内对总分的贡献不同，每个部分都有地板效应和天花板效应。特定的组成部分组合与预后相关，并且在不同TBI严重程度下，三个组成部分组合的预后价值高于总分。因此，我们建议在临床实践和预后研究中对三分法GCS进行多维度应用。
