背景和目的

静息心率（resting heart rate，RHR）与死亡之间的关联呈现出显著的地区与国家间异质性，而针对中国人群，目前尚缺乏来自全国性队列研究的可靠证据。既往研究普遍基于单次RHR的测量，而未考虑长期的累积效应。此外，现有风险评估模型同时考虑多个危险因素可以预测未来一段时间内心血管疾病（cardiovascular disease，CVD）的发病风险，但它和死亡之间的关联程度如何，目前缺乏相关研究。因此，本研究旨在评估RHR、累积静息心率（cumulative resting heart rate，cumRHR）及CVD预测风险对中国成年人死亡的独立及联合影响。

对象和方法

1992-2021年期间，中国动脉粥样硬化性心血管疾病风险预测研究（Prediction for Atherosclerotic Cardiovascular Disease Risk in China，China-PAR）对127,840名中国成年人进行了初始和多次随访调查。本研究纳入111,453名无重大慢性疾病者进行了RHR与CVD风险对死亡的影响分析，并以其中具有至少两次RHR测量的60,981人进行了cumRHR与CVD风险对死亡的影响分析。以2012-2015年调查作为基线，使用之前每两次连续检查的平均RHR与这两次检查之间的时间间隔的乘积之和，并进行5年加权作为cumRHR。根据三分位数将RHR和cumRHR分为低、中、高三组。使用China-PAR模型计算10年CVD发病风险，并以5%和10%为界分为低、中、高风险三组。本研究的终点事件为非意外死亡（ICD-10：A00-R99），包括CVD死亡（ICD-10：I00-I99）和非CVD死亡（ICD-10：A00-H95和J00-R99）。采用限制性立方样条（restricted cubic splines，RCSs）分析RHR（和cumRHR）及CVD预测风险与死亡的剂量-反应关系，采用Cox比例风险模型评估RHR（和cumRHR）和CVD预测风险与死亡相关的风险比（hazard ratio，HR）和95%置信区间（confidence interval，CI）以及累积死亡率，并进行性别和年龄的亚组分析以及敏感性分析。

结果

RHR与CVD预测风险对死亡的影响：在平均13.21年的随访期间，我们共观察到了12,270例非意外死亡，其中包括5,107例CVD死亡和7,163例非CVD死亡。RHR与死亡风险之间存在显著的剂量-反应关系，高RHR者的非意外死亡（1.42；95% CI：1.35，1.49）、CVD死亡（1.46；95% CI：1.36，1.57）和非CVD死亡（1.38；95% CI：1.29，1.47）的HR均高于低RHR者。10年CVD风险与非意外死亡和CVD死亡之间也存在显著的剂量-反应关系，与低10年CVD风险个体相比，高10年CVD风险个体的非意外死亡和CVD死亡的HR分别为1.67（95% CI：1.56，1.78）和3.38（95% CI：3.03，3.78）。随着10年CVD风险和RHR的升高，非意外死亡和CVD死亡的风险逐渐升高（所有P_趋势<0.01），高10年CVD风险且高RHR的个体非意外死亡（2.29；95% CI：2.09，2.52）和CVD死亡（5.26；95% CI：4.44，6.24）的HR最高。10年累积死亡率也显示出了同样的趋势（所有P_趋势<0.001）。

cumRHR与CVD预测风险对死亡的影响：在平均5.74年的随访期间，我们共观察到了3,141例非意外死亡，其中包括1,458例CVD死亡和1,683例非CVD死亡。cumRHR与死亡风险之间存在显著的剂量-反应关系，高cumRHR者的非意外死亡（1.35；95% CI：1.24，1.48）、CVD死亡（1.28；95% CI：1.13，1.45）和非CVD死亡（1.42；95% CI：1.26，1.60）的HR均高于低cumRHR者。随着10年CVD风险和cumRHR的升高，非意外死亡和CVD死亡的风险逐渐升高（所有P_趋势<0.01），高10年CVD风险且高cumRHR的个体非意外死亡（2.46；95% CI：1.92，3.16）和CVD死亡（4.58；95% CI：2.80，7.50）的HR最高。5年累积死亡率也显示出了同样的趋势（所有P_趋势<0.05）。

结论

我们的研究结果表明，RHR（和cumRHR）和CVD预测风险均与死亡独立相关，将二者结合起来可能有助于完善死亡风险分层。本研究强调了它们在个性化的零级预防和一级预防策略中的重要作用。

Background and objective

The association between resting heart rate (RHR) and mortality exhibits significant heterogeneity across diverse regions and countries. However, it is still a lack of reliable evidence from national cohorts in China. Previous studies have generally relied on a simple measurement of RHR without considering the long-term cumulative effects. Furthermore, while existing risk assessment models considering multiple risk factors simultaneously can predict incident cardiovascular disease (CVD) over time, research on the extent to which predicted CVD risk is associated with mortality remains scarce. Therefore, this study aims to evaluate the individual and combined impacts of RHR, cumulative resting heart rate (cumRHR), and predicted CVD risk on mortality among Chinese adults.

Subjects and methods

Between 1992 and 2021, the Prediction for Atherosclerotic Cardiovascular Disease Risk in China (China-PAR) study conducted initial and several follow-up surveys among 127,840 Chinese adults. 111,453 participants without major chronic diseases were included to estimate the impact of RHR and predicted CVD risk on mortality. Among them, 60,981 individuals with at least two RHR measurements were further analyzed to explore the relationship of cumRHR and predicted CVD risk on mortality. Using the 2012-2015 survey as the baseline, cumRHR was calculated as the 5-year weighted sum of the product of the average RHR from each 2 adjacent examinations and the time interval between them. Both RHR and cumRHR were categorized into low, intermediate, and high groups based on tertiles. The risk of CVD incidence was calculated using the China-PAR model, and divided into low, intermediate, and high-risk groups with cut-offs at 5% and 10%. The endpoint events in this study were non-accidental deaths (ICD-10: A00-R99), including CVD deaths (ICD-10: I00-I99) and non-CVD deaths (ICD-10: A00-H95 and J00-R99). Restricted cubic splines (RCSs) were used to examine the dose-response relationships of mortality with RHR (and cumRHR) and the predicted incident CVD risk. The Cox proportional hazards model was employed to assess the hazard ratios (HRs) and 95% confidence intervals (CIs) of mortality associated with RHR (and cumRHR) and the predicted risk of incident CVD, as well as cumulative mortality. Subgroup analyses by gender and age, along with sensitivity analyses, were also conducted.

Results

Impact of RHR and predicted CVD risk on mortality: During an average of 13.21 years of follow-up, we identified 12,270 non-accidental deaths, including 5,107 CVD deaths and 7,163 non-CVD deaths. A significant dose-response relationship was observed between RHR and mortality. High-RHR participants had higher HRs of non-accidental (1.42; 95% CI: 1.35, 1.49), CVD (1.46; 95% CI: 1.36, 1.57), and non-CVD mortality (1.38; 95% CI: 1.29, 1.47) than low-RHR ones, respectively. 10-year CVD risk also demonstrated a significant dose-response relationship with both non-accidental and CVD mortality. Compare to individuals with low 10-year CVD risk, those with high 10-year CVD risk demonstrated HRs of 1.67 (95% CI: 1.56, 1.78) and 3.38 (95% CI: 3.03, 3.78) for non-accidental and CVD mortality, respectively. The non-accidental and CVD mortality risks gradually increased with the elevation of 10-year CVD risk and RHR (all P for trend<0.01), showing the greatest HRs of non-accidental (2.29; 95% CI: 2.09, 2.52) and CVD mortality (5.26; 95% CI: 4.44, 6.24) among participants with both high 10-year CVD risk and high RHR. Similar trends were observed for 10-year cumulative mortality (all P for trend<0.001).

Impact of cumRHR and predicted CVD risk on mortality: During an average follow-up of 5.74 years, we observed 3,141 non-accidental deaths, including 1,458 CVD deaths and 1,683 non-CVD deaths. A significant dose-response relationship was observed between cumRHR and mortality. High-cumRHR individuals had higher HRs of non-accidental (1.35; 95% CI: 1.24, 1.48), CVD (1.28; 95% CI: 1.13, 1.45), and non-CVD mortality (1.42; 95% CI: 1.26, 1.60) compared to those with low cumRHR. The risks of non-accidental and CVD mortality gradually increased with rising 10-year CVD risk and cumRHR (all P for trend <0.01), with the highest HRs observed among individuals with both high 10-year CVD risk and high cumRHR for non-accidental (2.46; 95% CI: 1.92, 3.16) and CVD mortality (4.58; 95% CI: 2.80, 7.50). Similar trends were also observed for 5-year cumulative mortality (all P for trend <0.05).

Conclusion

Our findings demonstrated that RHR (and cumRHR) and predicted CVD risk are independently associated with mortality, and integrating them could potentially refine risk stratification of mortality. This study highlighted the important role of them in personalized strategies for primordial and primary prevention.