真实世界拉丁美洲国家中心源性休克评分预测性能的比较
本研究纳入872例心源性休克(CS)患者,计算SCAI、CARDSHOCK、IABP-Shock II、心源性休克评分、年龄-乳酸-肌酐评分、Get-With-The-Guidelines心力衰竭评分和急性失代偿性心力衰竭国家注册评分,用决策曲线分析评估净获益,用逻辑回归和Cox回归(逆概率删失加权法)构建AUC统计量预测院内死亡。逻辑回归分析中,整体队列CARDSHOCK评分AUC最高(c=0.666);急性心肌梗死相关CS(AMI-CS)中CARDSHOCK仍最高(0.68);非AMI-CS中仅SCAI(0.668)、CARDSHOCK(0.533)和IABP-SHOCK II(0.636)有统计学显著值。随时间分析,AUC出现显著差异,提示时间敏感成分影响死亡预测;CARDSHOCK最高(0.658),SCAI次之(0.622);AMI-CS中CARDSHOCK最高(0.671);非AMI-CS中SCAI最佳(0.642)。结论认为临床评分显示时间敏感的AUC,性能可能受时间和CS类型影响,理解时间影响可提供更好预测。
要点
- 整体队列逻辑回归中CARDSHOCK评分AUC最高(c=0.666),但所有评分仅中等至良好性能。
- AMI-CS患者中CARDSHOCK评分AUC最高(0.68);非AMI-CS中仅SCAI(0.668)、CARDSHOCK(0.533)、IABP-SHOCK II(0.636)有统计学显著值。
- 随时间分析AUC出现显著差异,提示时间敏感成分影响死亡预测;CARDSHOCK最高(0.658),SCAI次之(0.622)。
- AMI-CS随时间分析CARDSHOCK最高(0.671);非AMI-CS中SCAI最佳(0.642)。
- 临床评分性能受时间和CS类型影响,理解时间影响可能提供更好预测。
适用人群与场景
- 872例心源性休克(CS)患者,来自真实世界拉丁美洲国家。
研究设计与关键结果
- 研究类型为比较CS预后评分的真实世界研究,纳入872例CS患者,计算SCAI、CARDSHOCK、IABP-Shock II、心源性休克评分、年龄-乳酸-肌酐评分、Get-With-The-Guidelines心力衰竭评分和急性失代偿性心力衰竭国家注册评分,行决策曲线分析评估净获益。
- 逻辑回归预测院内死亡:整体队列CARDSHOCK评分AUC最高(c=0.666);AMI-CS中CARDSHOCK最高(0.68);非AMI-CS中仅SCAI(0.668)、CARDSHOCK(0.533)、IABP-SHOCK II(0.636)有统计学显著值。
- 随时间分析(Cox回归,逆概率删失加权法):AUC出现显著差异,提示时间敏感成分影响死亡预测;CARDSHOCK最高(0.658),SCAI次之(0.622);AMI-CS中CARDSHOCK最高(0.671);非AMI-CS中SCAI最佳(0.642)。
证据类型、等级与局限
- 证据类型:真实世界比较研究;局限:原文未报告局限性。
对床旁实践的改变
- 作者认为临床评分显示时间敏感的AUC,性能可能受时间和CS类型影响;理解时间影响可提供更好预测,并可能成为CS中有价值的工具。
提炼说明
- 本条依据论文摘要整理;摘要未涉及「诊断标准与关键阈值」「处置流程与药物剂量」「红旗征与不能漏的情况」「指南或共识出处」,故未列出这些栏目;全文请点「阅读原文」。
英文原文(摘要)· 点击展开
Background : Mortality in cardiogenic shock (CS) is up to 40%, and although risk scores have been proposed to stratify and assess mortality in CS, they have been shown to have inconsistent performance. The purpose was to compare CS prognostic scores and describe their performance in a real-world Latin American country. Methods : We included 872 patients with CS. The Society for Cardiovascular Angiography and Interventions (SCAI), CARDSHOCK, IABP-Shock II, Cardiogenic Shock Score, age-lactate-creatinine score, Get-With-The-Guidelines Heart Failure score, and Acute Decompensated Heart Failure National Registry scores were calculated. Decision curve analyses were performed to evaluate the net benefit of the different scoring systems. Logistic and Cox regression analyses were applied to construct area under the curve (AUC) statistics, this last one against time using the Inverse Probability of Censoring Weighting method, for in-hospital mortality prediction. Results: When logistic regression was applied, the scores had a moderate-good performance in the overall cohort that was higher AUC in the CARDSHOCK ( c = 0.666). In acute myocardial infarction-related CS (AMI-CS), CARDSHOCK still is the highest AUC (0.68). In non-AMI-CS only SCAI (0.668), CARDSHOCK (0.533), and IABP-SHOCK II (0.636) had statistically significant values. When analyzed over time, significant differences arose in the AUC, suggesting that a time-sensitive component influenced the prediction of mortality. The highest AUC was for the CARDSHOCK score (0.658), followed by SCAI (0.622). In AMI-CS-related, the highest AUC was for the CARDSHOCK score (0.671). In non-AMI-CS, SCAI was the best (0.642). Conclusions : Clinical scores show a time-sensitive AUC, suggesting that performance could be influenced by time and the type of CS. Understanding the temporal influence on the scores could provide a better prediction and be a valuable tool in CS.
参考来源
- Ortega-Hernández J, González-Pacheco H, Gopar-Nieto R, et al. COMPARISON OF THE PREDICTIVE PERFORMANCE OF CARDIOGENIC SHOCK SCORES IN A REAL-WORLD LATIN AMERICA COUNTRY. Shock. 2023. PMID: 36821419. doi:10.1097/SHK.0000000000002091
供执业医师抢救与诊疗参考,不替代临床判断;剂量请结合体重、肝肾功能和本院药品规格复核,以指南原文为准。