用评分法预测心肺复苏术后昏迷病人预后的临床研究

摘要

目的 建立心肺复苏自主循环恢复(CPRROSC)昏迷病人预后的评价方法 ,提高该类病人预后的预测能力.方法 分析文献,找出心肺复苏自主循环恢复昏迷病人预后的相关因素,赋予每个因素一定分值,建立CPRROSC预后评分法.用该评分法回顾性评价115例CPRROSC住院病人的预后,比较不同预后病人CPRROSC预后评分的差异,计算其对两种严重不良预后(死亡或植物状态)与其他类型预后区别能力的ROC曲线下面积.结果 5种不同预后(正常、轻度神经功能障碍、重度神经作者单位:518035深圳市第二人民医院急诊科(盂新科、赵志刚、吴光风、魏刚、刘德红、郑晓英、苏顺庭);中山大学附属第五医院妇产科(石少权)功能障碍、植物状态和死亡)病人CPRROSC预后评分比较.总的差异有统计学意义(F=65.91,P=0.000).其中正常组与神经功能轻度异常组、死亡组与植物状态组之间差异无统计学意义(3.52±3.03 vs 4.88±3.52,P=0.318;15.47±3.31 vs 14.04±3.84,P=0.108);其他各组之间相互比较差异均有统计学意义(植物状态组vs重度神经功能异常组为14.04±3.84 vs 10.70±3.30,P=0.011;其他各组之间比较,均为P=0.000).CPRROSC预后评分在8分以下对预后良好(正常或神经功能轻度异常)的病人区别能力最强;13分以上对预后严重不良的病人区别能力最强.CPRROSC预后评分对严重不良预后预测的ROC曲线下面积为0.950.结论 CPRROSC预后评分对病人严重不良预后具有较高预测和区别能力,可以作为心肺复苏后昏迷病人最终预后预测的评价工具.%Objective To exactly predict the prognoses of comatose patients with restoration of spontaneous circulation (ROSC) af-ter eardiopulmonary resuscitation (CPR) attempts, we adopted a predictive model based on summation score of multiple prognostic fac-tons. Methods We screen prognostic factors associated with survival after a resuscitation attempt by systematically reviewing published literature. Each factor included in the predictive model was assigned to a value. The total score of factors' values for a comatose patient was used to predict his/her outcome, so as what we call "CPRROSC predictive score". We retrospectively analyzed outcomes of 115 CPR-ROSC patients in coma using the predictive model. Score of patients with different outcomes was compared. Its predictive power for two categories of patients with poor outcomes and other patients was evaluated by calculating areas under ROC Curve. Results There were differences among CPRROSC predictive score of patients with five different outcomes (Good cerebral performance, moderate cerebral disa-bility, severe cerebral disability, vegetative state, and death) (F = 65.91 ,P = 0.000). However, There was no statistical difference be-tween patients with good cerebral performance and moderate cerebral disability (3.52±3.03 vs 4.88±3.52, P = 0.318), so it was with death and vegetative state(15.47±3.31 vs 14.04±3.84 P = 0.108). There were significant differences among patients with other out-comes. (vegetative state vs severe cerebral disability: 14.04±3.84 vs 10.70±3.30, P = 0.011). CPRROSC predictive Score was the most powerful indicator to predict patients with good cerebral performance when it was under 8 and patients with poor outcomes (death or vegetative state) when it was over 13. The area under ROC Curve for CPRROSC predictive score to predict outcomes of the alleged two kinds of patients was 0.95. Conclusion CPRROSC predictive score is more capable of exactly predicting the prognosis of patients with poor outcomes. It is also a useful modality to predict the final prognosis of comatose survivors after cardiopulmonary resuscitation.