4种风险评估方法预测晕厥短期不良事件的对比研究

Comparative study of four risk stratification tools for syncope

目的对预测晕厥患者短期内发生不良事件风险的4种方法进行对比研究。方法对2019年1月至2021年10月就诊于北京天坛医院的192例晕厥患者分别使用ROSE方法、ROSE-65方法、Boston方法、CSRS方法对照评估,并进行1个月随访。分别计算各方法的灵敏度、特异度、阳性预测值、阴性预测值、阳性似然比、阴性似然比,并对4种方法进行比较分析。结果随访1个月发现,对79例晕厥患者发生不良事件,ROSE、ROSE-65、Boston、CSRS 4种预测方法的特异度、阴性预测值、阴性似然比分别为84.07%、85.59%、0.241,79.65%、80.36%、0.350,85.84%、92.38%、0.118,79.65%、80.36%、0.350。ROSE、Boston方法的预测与发生不良事件一致性分析Kappa值分别为0.636、0.746,而2种方法间Kappa值为0.683。结论4种方法对非高风险晕厥患者的筛查具有不同的实践意义,应灵活使用以达到对患者危险分层快速评判的目的。

Objective To comparatively study four risk stratification tools for syncope in hospitalized patients.Methods From January 2019 to October 2021,192 patients with syncope admitted to our hospital were evaluated using the following four tools:Risk Stratification of Syncope in the Emergency Department(ROSE),ROSE-65,Boston Syncope Criteria,and Canadian Syncope Risk Score(CSRS).After 1 month of follow-up,the sensitivity,specificity,positive predictive value(PPV),negative predictive value(NPV),positive likelihood ratio(PLR),and negative likelihood ratio(NLR)were compared among the four tools.Results Seventy-nine patients had short-term serious outcomes.The specificity,NPV,and NLR values for the ROSE tool were 84.07%,85.59%,and 0.241;the corresponding values for the ROSE-65 tool were 79.65%,80.36%,and 0.350.Similarly,the specificity,NPV,and NLR values for the Boston Syncope Criteria were 85.84%,92.38%,and 0.118;the corresponding values for the CSRS tool were 79.65%,80.36%,and 0.350.Consistency analysis for the application of the ROSE and Boston Syncope Criteria tools with the result of follow-up adverse events yielded the Kappa values of 0.636 and 0.746 respectively,while the Kappa value obtained for the consistency between the two tools was 0.683.Conclusion All the four tools played important roles in the screening of non-high-risk patients.Therefore,it is important that physicians in the emergency department should quickly understand and evaluate potential risks for patients according to actual clinical situations.