应用SOAR评分预测急性缺血性卒中患者短期不良预后风险

目的:探讨国外SOAR评分对我国急性缺血性卒中(AIS)患者短期不良预后的预测价值。方法:前瞻性纳入我科2012年2月至2013年8月住院的AIS患者221例,收集临床相关基线资料并进行SOAR评分,分别以出院和卒中后3个月时功能残障和死亡定义为不良结局事件;通过受试者工作特征曲线下面积(AUC)判断SOAR评分的预测价值;使用Hosmer-Lemeshow法判断模型的拟合优度;应用Pearson相关分析评价实际与预期结局事件的关联程度。结果:出院和3个月时出现不良预后的患者分别为63例(28.5%)和71例(32.1%),死亡9例(4.1%)。SOAR评分在三个结局事件中的AUC分别为0.700、0.705、0.872;灵敏度分别为0.508、0.529、0.889;特异度分别为0.804、0.828、0.741;cut-off临界值为2分。Hosmer-Lemeshow法x2值分别为4.222、2.785、1.045(均P>0.05);Pearson相关系数分别为0.978、0.991、0.914(均P<0.05)。结论:SOAR评分对AIS患者短期不良结局发生风险预测价值较高,评分越高,出现不良预后的风险越大。

External Validation of SOAR Score to Predict Short Term Functional Outcomes after Acute Is-chemic Stroke in Chinese Population

ObjectiveTo assess the efficacy of the SOAR score, which was designed to predict short term risk af-ter strokes in Chinese acute ischemic stroke （AIS） patients. Methods：From December 2012 to August 2013, 221 AIS patients who were hospitalized in neurology department at our hospital were studied prospectively. The relat-ed baseline data were recorded and calculated by the SOAR score. The bad outcome events were defined as death, poor functional outcome at discharge and 3 months after the index stroke events. Model discrimination was quan-tified by calculating the area under the Receiver Operating Characteristic curve （AUC）, and the calibration was as-sessed using goodness of fit test and correlation coefficient. Results：The number of patients who had unfavorable outcome at discharge and 3 months were 63 （28.5%） and 71 （32.1%） respectively. Nine （4.1%） patients were dead at the end of the study. The AUCs of the three outcomes were 0.700, 0.705 and 0.872. The sensitivity was 0.508, 0.529 and 0.889 respectively and the specificity was 0.804, 0.828 and 0.741, respectively. The Cut-off value for three events was 2 points. The x2 of was 4.222, 2.785 and 1.045 （all P〈0.05）, and correlation coefficient was 0.978, 0.991 and 0.914 （all P〈0.05） respectively. Conclusion：SOAR score can predict short term functional outcomes in Chinese AIS people precisely and reliably. The higher scores correlate with the greater risk.