中文导读 高血压常被称为“沉默杀手”：全球逾十亿人受影响，却有过半成年人并不知道自己患病。谷歌研究院 2026 年 11 月 5 日刊于 Nature 的研究显示，消费级智能手表结合人工智能，或可在无需袖带校准的情况下发现高血压迹象。阅读时可留意 pressure、silent、optical、wrist、prompt、replace 等词，并注意作者强调这是机会性筛查，不能替代正式医学诊断。
正文 High blood pressure is often called a silent killer because many adults do not feel sick until a stroke or heart failure arrives . More than one billion people live with hypertension , yet over half of affected adults do not know they have the condition . On 5 November 2026, Google Research reported in Nature that artificial intelligence reading ordinary smartwatch signals may help close that awareness gap without first calibrating the watch against an arm cuff .
本段译文 高血压常被称为“沉默杀手”，因为许多成年人在中风或心力衰竭出现前并不觉得自己有病。全球逾十亿人患有高血压，但超过半数受影响的成年人并不知道自己有这一状况。2026年11月5日，谷歌研究院在《自然》报告称：用人工智能读取普通智能手表信号，或有助于缩小这种知晓差距——且不必先用臂式袖带给手表校准。
The system , named AI -PPG-ACC-HTN , does not try to print a single blood pressure number that jumps with every movement . Instead , it looks for a lasting hypertensive state by gathering pulse and motion data over several days . An optical sensor tracks how blood volume at the wrist changes with each heartbeat . A motion sensor captures tiny vibrations caused when the heart pushes blood into the aorta . Short waveform pieces are turned into mathematical features averaged across days to wash out noise , then combined with the wearer’s height — a rough proxy for arterial path length — to estimate hypertension probability
本段译文 名为 AI-PPG-ACC-HTN 的系统并不试图打印会随每一次动作跳动的单个血压读数。相反，它通过连续数日汇总脉搏与运动数据，寻找较为持久的高血压状态。光学传感器追踪手腕处血容量随每次心跳的变化；运动传感器捕捉心脏把血液推入主动脉时引起的细微振动。短段波形被转成数学特征，跨日平均以冲淡噪声，再与佩戴者身高（作为动脉路径长度的粗略代理）结合，估计高血压概率。
Researchers tested the idea in a prospective study at five U .S . sites . They enrolled 196 diverse adults who had no known cardiovascular disease and were not taking blood pressure medicine . The gold standard was 24-hour ambulatory monitoring . Over seven days of real-world wear , the smartwatch approach reached about 65.8 percent sensitivity , 90.0 percent specificity , and an 80.6 percent positive predictive value . First office screening , by comparison , showed about 55.3 percent sensitivity and a 77.8 percent positive predictive value at the same specificity .
本段译文 研究人员在美国五个地点开展前瞻性研究，纳入196名背景多样、无已知心血管疾病且未服用降压药的成年人。金标准是24小时动态血压监测。真实佩戴约七天后，智能手表方法的灵敏度约65.8%、特异度约90.0%、阳性预测值约80.6%。相比之下，初次诊室筛查在相同特异度下灵敏度约55.3%，阳性预测值约77.8%。
In a smaller group of 48 people whose office reading was already high , the watch confirmed sustained hypertension with about 78.4 percent sensitivity , 90.9 percent specificity , and a 96.7 percent positive predictive value . That performance was competitive with multi-day home cuff checks and stronger on some measures than a return office visit . Partners included Beth Israel Deaconess Medical Center and Harvard Medical School . Google says the findings helped inform a Health Guardian blood pressure trends feature , but the team stresses clear limits . The work is early feasibility research , not a device meant to replace clinical cuffs or to give a standalone medical diagnosis . Its intended role is opportunistic screening : a gentle prompt that at - risk individuals should seek formal assessment with established cuff methods before lasting cardiovascular damage sets in .
本段译文 在诊室读数已经偏高的48人小组中，手表确认持续性高血压的灵敏度约78.4%、特异度约90.9%、阳性预测值约96.7%。该表现可与多日家庭袖带复核竞争，并在部分指标上强于复诊诊室测量。合作方包括贝斯以色列女执事医疗中心与哈佛医学院。谷歌称这些发现帮助启发了 Health Guardian 血压趋势功能，但团队强调明确局限：这是早期可行性研究，不是用来替代临床袖带、也不是给出独立医学诊断的设备。其设想角色是机会性筛查——轻轻提示高风险者，应在持久心血管损害形成前，用既有袖带方法寻求正式评估。
新闻来源