负荷分布带心肺复苏期间的休克决策算法
该研究纳入896例心脏骤停患者负荷分布带(LDB)按压期间的5813段心电图,其中可电击1154段(303例)、不可电击4659段(841例)。LDB伪差较手动伪差振幅更大(5.5 mV vs 0.5 mV,p<0.001)、波形规律性更高(0.99 vs 0.88,p<0.001)、按压频率更低更稳定(80 min-1 vs 104.4 min-1)。用自适应滤波去除LDB伪差后设计机器学习算法进行休克决策,其准确度显著高于商用除颤器算法,敏感度/特异度92.1/96.8% vs 91.4/87.1%(p<0.001)。作者认为该机器学习算法在LDB按压期间实现了临床可靠的休克决策。
要点
- LDB按压期间心电图伪差振幅更大(5.5 mV vs 手动0.5 mV,p<0.001)、波形规律性更高(0.99 vs 0.88,p<0.001),去除更困难。
- LDB按压频率更低且更稳定,中位80 min-1(90%CI 79.9-82.9),手动伪差为104.4 min-1(90%CI 48.5-114.0)。
- 机器学习算法休克决策敏感度/特异度92.1/96.8%,商用除颤器算法为91.4/87.1%(p<0.001)。
- 数据集来自896例心脏骤停患者LDB按压期间的5813段心电图,可电击1154段(303例)、不可电击4659段(841例)。
适用人群与场景
- 心脏骤停并接受负荷分布带(LDB)心肺复苏的患者,研究数据集共896例、5813段心电图。
诊断标准与关键阈值
- 心电图段由共识标注为可电击(1154段,303例)或不可电击(4659段,841例)。
处置流程与药物剂量
- 使用自适应滤波去除LDB伪差,再经机器学习算法进行休克决策。
- 机器学习算法的休克决策准确度显著高于商用除颤器算法,敏感度/特异度92.1/96.8% vs 91.4/87.1%(p<0.001)。
研究设计与关键结果
- 研究类型为算法开发与性能比较研究,数据集为896例心脏骤停患者LDB按压期间的5813段心电图(可电击1154段/303例,不可电击4659段/841例)。
- LDB伪差与手动伪差比较:按压频率中位80 min-1(90%CI 79.9-82.9)vs 104.4 min-1(90%CI 48.5-114.0),LDB更低且更稳定。
- 振幅5.5 mV(0.8-23.4)vs 0.5 mV(0.1-2.2),LDB更大(p<0.001);波形规律性(Pearson相关系数)0.99(0.78-1.0)vs 0.88(0.55-0.98),LDB更高(p<0.001)。
- 休克决策敏感度/特异度:机器学习算法92.1/96.8% vs 除颤器算法91.4/87.1%,机器学习算法显著更高(p<0.001)。
证据类型、等级与局限
- 证据类型:算法开发与性能比较研究。
- 局限:原文未报告局限性。
对床旁实践的改变
- 作者认为,机器学习算法在LDB按压期间实现了临床可靠的休克决策,其敏感度/特异度92.1/96.8%高于商用除颤器算法的91.4/87.1%(p<0.001)。
提炼说明
- 本条依据论文摘要整理;摘要未涉及「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
AIM: Chest compressions delivered by a load distributing band (LDB) induce artefacts in the electrocardiogram. These artefacts alter shock decisions in defibrillators. The aim of this study was to demonstrate the first reliable shock decision algorithm during LDB compressions.
METHODS: The study dataset comprised 5813 electrocardiogram segments from 896 cardiac arrest patients during LDB compressions. Electrocardiogram segments were annotated by consensus as shockable (1154, 303 patients) or nonshockable (4659, 841 patients). Segments during asystole were used to characterize the LDB artefact and to compare its characteristics to those of manual artefacts from other datasets. LDB artefacts were removed using adaptive filters. A machine learning algorithm was designed for the shock decision after filtering, and its performance was compared to that of a commercial defibrillator's algorithm.
RESULTS: Median (90% confidence interval) compression frequencies were lower and more stable for the LDB than for the manual artefact, 80 min-1 (79.9-82.9) vs. 104.4 min-1 (48.5-114.0). The amplitude and waveform regularity (Pearson's correlation coefficient) were larger for the LDB artefact, with 5.5 mV (0.8-23.4) vs. 0.5 mV (0.1-2.2) (p < 0.001) and 0.99 (0.78-1.0) vs. 0.88 (0.55-0.98) (p < 0.001). The shock decision accuracy was significantly higher for the machine learning algorithm than for the defibrillator algorithm, with sensitivity/specificity pairs of 92.1/96.8% (machine learning) vs. 91.4/87.1% (defibrillator) (p < 0.001).
CONCLUSION: Compared to other cardiopulmonary resuscitation artefacts, removing the LDB artefact was challenging due to larger amplitudes and lower compression frequencies. The machine learning algorithm achieved clinically reliable shock decisions during LDB compressions.
参考来源
- Isasi I, Irusta U, Aramendi E, et al. Shock decision algorithm for use during load distributing band cardiopulmonary resuscitation. Resuscitation. 2021. PMID: 34098032. doi:10.1016/j.resuscitation.2021.05.028
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。