This year, ≈400 000 Americans will have a sudden cardiac arrest (SCA). Despite a sophisticated chain of survival that involves cardiopulmonary resuscitation, use of automated external defibrillators, and transport for advanced hospital care, ≈10% will survive. The vast majority will experience a sudden cardiac death (SCD). The primary prevention implantable cardioverter defibrillator (ICD, largely for those with left ventricular ejection fraction <35%) can prevent SCD with a therapeutic shock that occurs within seconds. However, the ≈150 000 that will be implanted this year will deliver appropriate shocks in 3% to 5%,¹ saving an estimated average number of 6000 lives per year. These gradually diminishing returns from the ICD can be attributed to “getting with the guidelines” and a steady stream of successful new treatments for heart failure. The field continues to work on refining SCD risk stratification, especially for high-risk individuals with left ventricular ejection fraction >35%, but these developments will take time, running the gamut of new clinical trials and revisions to clinical guidelines. In the meantime, could we develop a new strategy for SCD prevention beyond automated external defibrillators and ICDs, with a faster turnaround time?

We have proposed a complementary approach for SCD prevention interposed between long-term prevention (ICDs) and SCA resuscitation (automated external defibrillators), that we call “near-term prevention.”² Near-term prevention is predicated on harnessing warning symptoms that occur before imminent SCA. Warning signs have long been recognized as harbingers of SCA, but we reported for the first time that an individual’s response to their warning symptoms can be associated with survival from SCA.³ This analysis from the Oregon Sudden Unexpected Death Study showed that of 839 individuals experiencing SCA (age 35–65 years; 75% men), half had warning symptoms (50% of men and 53% of women). For the majority of individuals, a 911 call was made after collapse from SCA. However, the minority (19%) of individuals for whom 911 was contacted early on the basis of warning symptoms, had a 5- to 6-fold higher rate of survival after SCA resuscitation. On the basis of the emergency medical services literature, this was likely due to earlier initiation of resuscitation by first responders.

It was abundantly clear that an approach that would require residents of the general population to call 911 for symptoms of chest pain or dyspnea was not ready for prime time. Because these symptoms overlap with other conditions, only a small minority of 911 calls would be for SCA, a quick recipe for emergency medical services systems grinding to a halt due to false-positive calls. There was need for a study conducted with a comparison group, with the goal of teasing out the symptoms most likely to be associated with SCA. We conducted a study in Ventura County, CA (discovery, population ≈850 000 residents) and Multnomah County, OR (replication, population ≈800 000).⁴ Two groups of cases (ages 18–85) of witnessed SCA (2015–2021) and corresponding control groups of individuals that did not experience SCA were identified from 911 emergency calls. We compared symptoms reported to emergency medical services before imminent SCA, with symptoms reported for non-SCA 911 calls. Exactly half of 823 SCA cases in the discovery group called in at least 1 warning symptom. When compared with the control group (n=1171), the vast majority of symptom types were not associated with SCA, with no patterns identified for clusters of symptoms. However, SCA cases were significantly more likely to have dyspnea (41% versus 22%), chest pain (33% versus 25%), diaphoresis (12% versus 8%), and seizure-like activity (11% versus 7%). Warning symptoms were sex-specific, with chest pain, dyspnea, and diaphoresis being significantly associated in men (area under the receiving operator characteristic curve=0.644) and only dyspnea significantly associated with SCA in women (area under the receiving operator characteristic curve=0.686). Very similar findings were observed in the Oregon replication population. This study was able to exclude the majority of symptom types (eg, nausea, weakness, dizziness, palpitations) from future consideration. We concluded that warning symptoms of SCA are sex-specific and have the potential to predict imminent SCA but will need to be enhanced with additional features before deployment for near-term prevention of SCD. These additional features could include individual risk stratifiers such as the clinical profile, ECG during symptoms or measured biomarkers.

As the predictive power of near-term prevention algorithms reaches threshold, how will such approaches be deployed? The first setting is likely to be the general population (Figure). We have previously highlighted how smart devices enabled by artificial intelligence algorithms have the potential to perform rapid triage in the acute setting. On the basis of an individual’s score on the algorithm, they may be prompted by a smart device to urgently seek care instead of ignoring their warning symptoms, thus facilitating near-term SCA prevention. Smart devices also have the potential to intervene at the penultimate moment. Approximately 75% to 80% of SCA events occur at home, and if these are unrecognized, there is no hope of survival. Chan et al⁵ have successfully used “agonal breathing,” an underappreciated diagnostic sign of SCA. Using a “smart speaker,” they captured sounds of agonal breathing during SCA events and trained a smart vector classifier to distinguish these from sounds produced by other individuals during sleep. Their smart speaker could successfully distinguish SCA agonal breathing from sleep sounds. Their findings are a readily applicable example of how remote, noncontact, passive detection of unwitnessed SCA could enable a rapid emergency medical services response and facilitate survival. A second setting in which near-term prevention could be deployed is the emergency department. It can often be very challenging to successfully distinguish symptomatic individuals who should be admitted for observation and further investigation versus those who should be discharged with clinical follow-up.

Figure. Near-term prevention of sudden cardiac death. Among high-risk individuals who receive primary prevention ICDs, SCD can be averted by an appropriate shock. Among others who do not receive ICDs, smart algorithms based on warning symptoms and additional features could have the potential to recognize high risk, triage them for management, and avert imminent SCA. For asymptomatic individuals or others in whom risk goes unrecognized, smart devices have the potential to detect unwitnessed SCAs early and facilitate survival. AED indicates automated external defibrillator; ICD, implantable cardioverter defibrillator; SCA, sudden cardiac arrest; and SCD, sudden cardiac death.

In summary, automated external defibrillators and ICDs continue to have an effect on SCD prevention, but the effectiveness of these important modalities may have plateaued. Further investigation and clinical trials could produce enhancements, but this is likely to be a long road. Enabled by the ongoing surge in artificial intelligence, it is likely that near-term SCD prevention could be deployed faster. Near-term SCD prevention algorithms and smart devices will also require evaluation in clinical trials before deployment in the community or emergency department.

The author expresses deep appreciation for his colleagues at the Center for Cardiac Arrest Prevention, Cedars-Sinai Smidt Heart Institute and emergency medical services coinvestigators/collaborators in both Multnomah County, OR, and Ventura County, CA, whose efforts translated into the findings summarized in these pages; and to the residents of these communities without whom these studies could not have been conducted.

Funded by National Institutes of Health, National Heart Lung and Blood Institute grants R01HL145675 and R01HL147358 to Dr Chugh. Dr Chugh holds the Pauline and Harold Price Chair in Cardiac Electrophysiology at Cedars-Sinai.

Disclosures None.

The American Heart Association celebrates its 100th anniversary in 2024. This article is part of a series across the entire AHA Journal portfolio written by international thought leaders on the past, present, and future of cardiovascular and cerebrovascular research and care. To explore the full Centennial Collection, visit https://www.ahajournals.org/centennial

The opinions expressed in this article are not necessarily those of the editors or of the American Heart Association.

For Sources of Funding and Disclosures, see page 1061.

Circulation is available at www.ahajournals.org/journal/circ

今年，约 40 万美国人将发生心脏骤停 (SCA)。尽管存在复杂的生存链，涉及心肺复苏、使用自动体外除颤器以及前往高级医院护理的运输，但大约有 10% 的人能够存活下来。绝大多数人会经历心源性猝死（SCD）。一级预防植入式心脏复律除颤器（ICD，主要针对左心室射血分数 <35% 的患者）可以通过在几秒钟内发生的治疗性休克来预防 SCD。然而，今年将植入的约 150,000 个将产生 3% 至 5% 的适当冲击，¹估计每年平均挽救 6000 人的生命。 ICD 的回报逐渐递减可归因于“遵守指南”和源源不断的成功的心力衰竭新疗法。该领域继续致力于完善 SCD 风险分层，特别是对于左心室射血分数 >35% 的高危个体，但这些进展需要时间，需要进行一系列新的临床试验和临床指南的修订。与此同时，除了自动体外除颤器和 ICD 之外，我们是否可以制定一种新的 SCD 预防策略，并且周转时间更快？

我们提出了一种介于长期预防 (ICD) 和 SCA 复苏（自动体外除颤器）之间的 SCD 预防补充方法，我们称之为“近期预防”。²近期预防的基础是利用即将发生 SCA 之前出现的警告症状。警告信号长期以来一直被认为是 SCA 的先兆，但我们首次报告称，个体对其警告症状的反应可能与 SCA 的生存有关。³俄勒冈州意外猝死研究的分析显示，在 839 名经历 SCA 的人（年龄 35-65 岁；75% 男性）中，一半有警告症状（50% 男性和 53% 女性）。对于大多数人来说，在 SCA 晕倒后会拨打 911 电话。然而，少数 (19%) 因出现警告症状而提前联系 911 的人，在 SCA 复苏后的存活率高出 5 至 6 倍。根据紧急医疗服务文献，这可能是由于急救人员较早开始复苏所致。

很明显，要求普通居民因胸痛或呼吸困难症状拨打 911 的方法还没有准备好。由于这些症状与其他情况重叠，因此只有一小部分 911 电话是针对 SCA 的，这是紧急医疗服务系统因误报电话而陷入瘫痪的快速方法。需要对对照组进行一项研究，目的是找出最有可能与 SCA 相关的症状。我们在加利福尼亚州文图拉县（发现，人口约 85 万居民）和俄勒冈州摩特诺玛县（复制，人口约 80 万）进行了一项研究。⁴从 911 紧急呼叫中确定了两组目睹过 SCA 的病例（18-85 岁）（2015-2021 年）和相应的对照组（未经历过 SCA）。我们将即将发生 SCA 之前向紧急医疗服务部门报告的症状与非 SCA 911 呼叫报告的症状进行了比较。在发现组的 823 例 SCA 病例中，恰好有一半出现了至少 1 种警告症状。与对照组 (n=1171) 相比，绝大多数症状类型与 SCA 无关，也没有确定症状群的模式。然而，SCA 病例明显更有可能出现呼吸困难（41% 对 22%）、胸痛（33% 对 25%）、出汗（12% 对 8%）和癫痫样活动（11% 对 7%） 。警告症状具有性别特异性，其中胸痛、呼吸困难和出汗与男性显着相关（接收操作者特征曲线下面积=0.644），只有呼吸困难与女性 SCA 显着相关（接收操作者特征曲线下面积=0.686） ）。在俄勒冈州的复制群体中也观察到了非常相似的结果。这项研究能够从未来的考虑中排除大多数症状类型（例如恶心、虚弱、头晕、心悸）。我们得出的结论是，SCA 的警告症状具有性别特异性，并且有可能预测即将发生的 SCA，但在部署用于近期预防 SCD 之前需要通过其他功能进行增强。这些附加特征可能包括个人风险分层，例如临床概况、症状期间的心电图或测量的生物标志物。

随着近期预防算法的预测能力达到阈值，这些方法将如何部署？第一个设定很可能是普通人群（图）。我们之前强调过人工智能算法支持的智能设备如何有潜力在紧急情况下进行快速分类。根据个人在算法上的得分，智能设备可能会提示他们紧急寻求护理，而不是忽视他们的警告症状，从而促进近期的 SCA 预防。智能设备也有可能在倒数第二时刻进行干预。大约 75% 至 80% 的 SCA 事件发生在家里，如果这些事件未被识别，则没有生存的希望。 Chan 等人⁵成功地使用了“临终呼吸”，这是一种未被充分认识的 SCA 诊断标志。他们使用“智能扬声器”捕捉 SCA 事件期间痛苦呼吸的声音，并训练智能矢量分类器将这些声音与其他人在睡眠期间产生的声音区分开来。他们的智能扬声器可以成功区分 SCA 临终呼吸和睡眠声音。他们的研究结果是一个易于应用的例子，说明远程、非接触、被动检测无人目击的 SCA 如何能够实现快速紧急医疗服务响应并促进生存。可以部署近期预防的第二种环境是急诊科。成功区分应入院观察和进一步检查的有症状个体与应出院进行临床随访的个体通常非常具有挑战性。

数字。 近期预防心源性猝死。在接受初级预防 ICD 的高危人群中，适当的电击可以避免 SCD。对于其他没有接受 ICD 的人来说，基于警告症状和附加功能的智能算法可能有潜力识别高风险，对他们进行分类管理，并避免即将发生的 SCA。对于无症状的个人或其他风险未被识别的人来说，智能设备有可能尽早发现无人目击的 SCA 并促进生存。 AED表示自动体外除颤器； ICD，植入式心脏复律除颤器； SCA，心脏骤停；和 SCD，心源性猝死。

总之，自动体外除颤器和 ICD 继续对 SCD 预防产生影响，但这些重要方法的有效性可能已趋于稳定。进一步的研究和临床试验可能会带来改进，但这可能是一条漫长的道路。在人工智能不断激增的推动下，近期的 SCD 预防措施可能会更快部署。近期的 SCD 预防算法和智能设备在部署到社区或急诊室之前也需要进行临床试验评估。

作者对他在心脏骤停预防中心、雪松-西奈·斯密特心脏研究所的同事以及俄勒冈州摩特诺玛县和加利福尼亚州文图拉县的紧急医疗服务共同调查员/合作者表示深深的谢意，他们的努力转化为总结的研究结果这些页面；以及这些社区的居民，没有他们，这些研究就不可能进行。

在美国国立卫生研究院的资助下，国家心肺和血液研究所向 Chugh 博士授予 R01HL145675 和 R01HL147358。 Chugh 博士担任 Cedars-Sinai 心脏电生理学领域的 Pauline 和 Harold Price 主席。

披露无。

美国心脏协会将于 2024 年庆祝成立 100 周年。本文是国际思想领袖撰写的整个 AHA 期刊系列文章的一部分，内容涉及心脑血管研究和护理的过去、现在和未来。要探索完整的百年纪念收藏，请访问 https://www.ahajournals.org/centennial

本文表达的观点不一定代表编辑或美国心脏协会的观点。

有关资金来源和披露信息，请参阅第 1061 页。

流通量可在 www.ahajournals.org/journal/circ 上获取