Chinese General Practice ›› 2026, Vol. 29 ›› Issue (03): 299-310.DOI: 10.12114/j.issn.1007-9572.2025.0109

Special Issue: 心血管最新文章合辑

• Article • Previous Articles    

Effect of Blood Pressure Trajectory on Cardiovascular Disease in Population with High-normal Blood Pressure: a Study Based on the Kailuan Cohort

  

  1. 1. School of Public Health, North China University of Science and Technology, Tangshan 063000, China
    2. Department of Cardiology, Kailuan General Hospital, Tangshan 063000, China
    3. School of Clinical Medicine, North China University of Science and Technology, Tangshan 063000, China
    4. Department of Neurosurgery, Affiliated Hospital of North China University of Science and Technology, Tangshan 063000, China
  • Received:2025-04-21 Revised:2025-10-14 Published:2026-01-20 Online:2025-12-11
  • Contact: YANG Peng, WU Shouling, LI Yun
  • About author:

    LU Yanqiu and WU Yuntao are co-first authors

正常高值血压人群血压轨迹对心血管疾病发病风险的影响:基于开滦队列研究

  

  1. 1.063000 河北省唐山市,华北理工大学公共卫生学院
    2.063000 河北省唐山市,开滦总医院心内科
    3.063000 河北省唐山市,华北理工大学临床医学院
    4.063000 河北省唐山市,华北理工大学附属医院骨科
  • 通讯作者: 杨鹏, 吴寿岭, 李云
  • 作者简介:

    鹿妍秋和吴云涛为共同第一作者

    作者贡献:

    鹿妍秋、吴云涛、杨鹏、吴寿岭、李云提出研究思路,设计研究方案;鹿妍秋、吴云涛、刘少鹏、林海颖、邓惠友、武英、黄喆负责研究的构思与设计,研究的实施,进行数据的收集与整理,统计学处理,图、表的绘制与展示,负责论文起草撰写;杨鹏、吴寿岭、李云负责最终版本修订,文章的质量控制与审查,对文章整体负责,监督管理。

Abstract:

Background

Cardiovascular disease (CVD) is a leading cause of premature death and rising healthcare costs worldwide. An estimated 330 million people in China are affected by CVD, posing a substantial medical and financial burden to the nation. Among Chinese adults aged 18 and over without a prior hypertension diagnosis, 50.9% have high-normal blood pressure (BP). High-normal BP is associated with an increased risk of CVD incidence, and its effective control could prevent over 10% of CVD cases.

Objective

This study aimed to identify distinct BP trajectory patterns in individuals with high-normal BP using group-based trajectory modeling, investigate the impact of these trajectories on CVD risk, and provide a scientific basis for developing BP management and CVD prevention strategies for this population.

Methods

Participants from the Kailuan Study who underwent health examinations between 2006 and 2012 and were identified as having high-normal BP at their first examination were enrolled. Baseline data were collected via questionnaires, anthropometric measurements, and laboratory tests. Participants were followed until the first occurrence of a CVD event [myocardial infarction (MI) or stroke] or until December 31, 2022. Group-based trajectory modeling (GBTM) was used to model systolic BP (SBP) and diastolic BP (DBP) trajectories. Kaplan-Meier survival curves and Log-rank tests were used to compare cumulative CVD incidence across trajectory groups. Multivariable Cox proportional hazards regression models, or weighted Cox regression models when the proportional hazards assumption was violated, were used to analyze CVD risk across different BP trajectory groups. Multiplicative interactions between sex/age and BP trajectories were tested using Cox models, followed by stratified analyses. Sensitivity analyses were conducted to verify the robustness of the findings.

Results

A total of 21 745 subjects were ultimately included, with a mean age of (54.0±11.4) years; 17 556 (80.74%) were male and 4 289 (19.26%) were female. GBTM results identified four distinct SBP trajectories, each characterised by its SBP range and pattern of change over follow-up (i.e., stable, declining, or increasing). Among 7 088 subjects maintained stable SBP around 115 mmHg (1 mmHg=0.133 kPa), termed the "normal-high blood pressure decline group"; 11 662 subjects maintained stable SBP around 130 mmHg, termed the "normal-high blood pressure stable group"; 1 710 subjects exhibited a gradual increase from 129 mmHg to approximately 160 mmHg during the trajectory period, termed the "normal high blood pressure ascending cohort"; 1 285 subjects saw their SBP rise to around 158 mmHg before declining below 140 mmHg, termed the "normal high blood pressure ascending-descending cohort". DBP identified four distinct trajectories, each characterised by its DBP range and pattern of change over follow-up (i.e., stable, declining, or increasing). Among 4 856 subjects, DBP remained stable around 75 mmHg, termed the "normal-high blood pressure declining group"; 13 668 subjects exhibited stable DBP around 83 mmHg, termed the "normal-high blood pressure stable group"; 1 640 subjects exhibited a gradual increase in DBP from 82 mmHg to approximately 100 mmHg during the trajectory period, termed the "normal high blood pressure rising group"; 1 581 subjects experienced an increase in SBP to around 98 mmHg before declining below 90 mmHg, termed the "normal high blood pressure rising-falling group". Participants across different SBP trajectory groups were comparable in age, sex, BMI, educational attainment, smoking, alcohol consumption, physical activity, SBP, DBP, low-density lipoprotein cholesterol (LDL-C), triglycerides (TG), total cholesterol (TC), fasting blood glucose (FBG), high-sensitivity C-reactive protein, estimated glomerular filtration rate (eGFR), isolated systolic hypertension (ISH), isolated diastolic hypertension (IDH), hypertension, diabetes, antihypertensive medication use, and antidiabetic medication use were compared across different DBP trajectory groups. Differences were statistically significant (P<0.05). Comparisons of age, gender, BMI, educational attainment, smoking, alcohol consumption, salt preference, SBP, DBP, HDL-C, LDL-C, TG, TC, FBG, high-sensitivity C-reactive protein, eGFR, ISH, IDH, hypertension, diabetes mellitus, and antihypertensive medication use showed statistically significant differences (P<0.05). The mean follow-up duration was (9.43±1.94) years, during which 1 429 CVD events occurred (259 MI, 1 170 strokes, and 20 concurrent events). The SBP trajectory subgroups (normal high blood pressure decline, stable, increase, and increase-decline) recorded 274, 794, 191, and 170 new CVD cases respectively. The DBP trajectory groups showed 227, 881, 163, and 158 new CVD cases in the normal-high blood pressure decline, stable, increase, and increase-decline groups respectively. Survival curves depicting cumulative CVD incidence were plotted for study subjects. Log-rank test results indicated statistically significant differences in cumulative CVD incidence between different SBP trajectory groups and DBP trajectory groups (χ2=275.39, 90.69; P<0.001). Using Cox proportional hazards regression models, the hazard ratios (HR) for CVD occurrence across different SBP trajectory groups (95%CI) were calculated, with the normal-high blood pressure decline group serving as the reference. The HR for CVD occurrence in the normal-high blood pressure stable group, normal-high blood pressure increase group, and normal-high blood pressure increase-decrease group were 1.41 (1.22-1.62), 1.92 (1.58-2.33), and 2.24 (1.84-2.74), respectively. The HR (95%CI) for stroke occurrence were 1.46 (1.25-1.71), 2.04 (1.65-2.53) and 2.37 (1.90-2.96), respectively. The HR (95%CI) for MI occurrence were 1.25 (0.92-1.72), 1.42 (0.90-2.23) and 1.81 (1.14-2.86), respectively. Using weighted multivariable Cox proportional hazards regression models unconstrained by the proportional hazards assumption, the aHR (95%CI) for CVD occurrence was calculated with the normal-to-high blood pressure decrease group as the reference. The aHR (95%CI) for CVD occurrence in the normal-to-high blood pressure stable group, the normal-to-high blood pressure increase group, and the normal-to-high blood pressure increase-decrease group were 1.43 (1.12-1.82), 2.59 (1.62-4.13), and 2.11 (1.40-3.17), respectively. The aHR (95%CI) for stroke occurrence were 1.45 (1.11-1.71), 2.95 (1.75-4.97) and 2.34 (1.48-3.71), respectively. The aHR (95%CI) for MI occurrence were 1.34 (0.76-2.34), 1.17 (0.62-2.19) and 1.32 (0.69-2.55), respectively. Stratified analysis indicated that only the SBP trajectory showed a statistically significant interaction with gender and age (Pinteraction<0.05), whereas the DBP trajectory did not exhibit a statistically significant interaction with gender and age (Pinteraction>0.05). In the male cohort, using the SBP trajectory normal-high blood pressure decreasing group as the reference group, the HR (95%CI) for CVD occurrence in the normal-high blood pressure stable group, normal-high blood pressure increasing group, and normal-high blood pressure increasing-decreasing group were 1.42 (1.20-1.67), 1.93 (1.53-2.42), and 2.30 (1.82-2.90), respectively; In the female cohort, using the SBP trajectory normal-high blood pressure decline group as the reference group, the HR (95%CI) for CVD occurrence in the stable, rising, and rising-decline groups were 1.80 (1.14-2.85), 2.91 (1.62-5.21), and 2.79 (1.43-5.43), respectively. Among individuals aged <60 years with prehypertension, using the SBP trajectory prehypertension decline group as reference, the HR (95%CI) for CVD occurrence were 1.48 (1.22-1.78), 2.00 (1.46-2.71), and 3.01 (2.24-4.06), respectively. Among normotensive individuals aged ≥60 years, with the declining SBP trajectory group as reference, CVD HR (95%CI) was 1.35 (1.01-1.79), 1.94 (1.40-2.69), and 1.91 (1.35-2.69), respectively.

Conclusion

Distinct BP trajectory patterns are associated with varying levels of CVD risk in individuals with high-normal BP. For SBP, trajectories characterized by high levels and large fluctuations carried the highest CVD risk. For DBP, trajectories with an ascending trend and high baseline DBP posed the greatest risk. Even individuals whose BP remains stable within the high-normal range (120-139/80-89 mmHg) without major fluctuations still face an elevated risk of CVD.

Key words: Cardiovascular disease, High-normal blood pressure, Blood pressure trajectory, Kailuan cohort study, Group-based trajectory modeling, Cox proportional hazards regression model

摘要:

背景

心血管疾病(CVD)是全球过早死亡和医疗费用上涨的主要原因。据估计,中国约有3.3亿人受到CVD的影响,给国家带来了巨大的医疗和财政负担。中国18岁以上未被诊断过高血压的成年人中,有50.9%的人处于正常高值血压,正常高值血压与CVD发病风险增加有关,有效控制正常高值血压可预防10%以上的CVD病例。

目的

本研究采用群组轨迹模型(GBTM)探索正常高值血压人群的血压轨迹变化模式,探究正常高值血压人群血压轨迹对CVD发病风险的影响,为制订正常高值血压人群血压管理和CVD预防策略提供科学依据。

方法

选取参加开滦研究2006—2012年度健康体检且首次参加体检时为正常高值血压者为研究对象,通过问卷调查、人体测量学指标检查及实验室生化检测收集基线资料。对患者进行随访,随访结局事件为首次发生CVD事件[心肌梗死(MI)和脑卒中],随访终止时间为2022-12-31。采用GBTM拟合研究对象的收缩压(SBP)和舒张压(DBP)的轨迹模型。采用Kaplan-Meier生存曲线分析不同轨迹组CVD累积发病率,Log-rank检验比较组间差异。采用多因素Cox比例风险回归模型或不受比例风险假设约束的加权多因素Cox比例风险回归模型分析不同血压轨迹组发生CVD事件的风险。采用Cox比例风险回归模型分析性别、年龄与血压轨迹之间存在的潜在相乘交互作用并进行分层分析。进行敏感性分析验证结果的稳健性。

结果

最终纳入21 745名研究对象,平均年龄(54.0±11.4)岁,男17 556名(80.74%),女4 289名(19.26%)。GBTM结果显示,SBP确定了4条不同的轨迹,每条轨迹均根据其SBP范围和随体检时间变化的模式(即稳定、下降和上升)进行标记,7 088名研究对象的SBP稳定在115 mmHg(1 mmHg=0.133 kPa)左右,称为"正常高值血压下降组",11 662名研究对象的SBP稳定在130 mmHg左右,称为"正常高值血压稳定组",1 710名研究对象在轨迹期从129 mmHg逐渐增长到160 mmHg左右,称为"正常高值血压上升组",1 285名研究对象的SBP增长到158 mmHg左右后再下降到140 mmHg以下,称为"正常高值血压上升-下降组"。DBP确定了4条不同的轨迹,每条轨迹都根据其DBP范围和随体检时间变化的模式(即稳定、下降和上升)进行标记,4 856名研究对象的DBP稳定在75 mmHg左右,称为"正常高值血压下降组",13 668名研究对象的DBP稳定在83 mmHg左右,称为"正常高值血压稳定组",1 640名研究对象在轨迹期DBP从82 mmHg逐渐增长到100 mmHg左右,称为"正常高值血压上升组",1 581名研究对象的SBP增长到98 mmHg左右后再下降到90 mmHg以下,称为"正常高值血压上升-下降组"。不同SBP轨迹组中研究对象年龄、性别、BMI、受教育程度、吸烟、饮酒、体育锻炼、SBP、DBP、低密度脂蛋白胆固醇(LDL-C)、三酰甘油(TG)、总胆固醇(TC)、空腹血糖(FBG)、超敏C反应蛋白、估算肾小球滤过率(eGFR)、单纯收缩期高血压(ISH)、单纯舒张期高血压(IDH)、高血压、糖尿病、服用降压药、服用降糖药比例比较,差异有统计学意义(P<0.05)。不同DBP轨迹组中研究对象年龄、性别、BMI、受教育程度、吸烟、饮酒、喜盐、SBP、DBP、高密度脂蛋白胆固醇(HDL-C)、LDL-C、TG、TC、FBG、超敏C反应蛋白、eGFR、ISH、IDH、高血压、糖尿病、服用降压药比例比较,差异有统计学意义(P<0.05)。患者平均随访时间(9.43±1.94)年,共计发生CVD1 429例(259例MI,1 170例脑卒中,两者同时发生20例)。SBP轨迹分组正常高值血压下降组、正常高值血压稳定组、正常高值血压上升组和正常高值血压上升-下降组新发CVD病例分别为274例、794例、191例和170例;DBP轨迹分组正常高值血压下降组、正常高值血压稳定组、正常高值血压上升组和正常高值血压上升-下降组CVD新发病例分别为227例、881例、163例和158例。绘制研究对象CVD累积发病率生存曲线,Log-rank检验结果显示不同SBP轨迹组、DBP轨迹组CVD累积发病率比较,差异有统计学意义(χ2=275.39、90.69,P<0.001)。采用Cox比例风险回归模型计算不同SBP轨迹组发生CVD的HR(95%CI),以正常高值血压下降组作为参照组,正常高值血压稳定组、正常高值血压上升组和正常高值血压上升-下降组发生CVD的HR(95%CI)分别为1.41(1.22~1.62)、1.92(1.58~2.33)和2.24(1.84~2.74),发生脑卒中的HR(95%CI)分别为1.46(1.25~1.71)、2.04(1.65~2.53)和2.37(1.90~2.96),发生MI的HR(95%CI)分别为1.25(0.92~1.72)、1.42(0.90~2.23)和1.81(1.14~2.86)。使用不受比例风险假设约束的加权多因素Cox比例风险回归模型计算发生CVD的aHR(95%CI),以正常高值血压下降组作为参照组,正常高值血压稳定组、正常高值血压上升组和正常高值血压上升-下降组发生CVD的aHR(95%CI)分别为1.43(1.12~1.82)、2.59(1.62~4.13)和2.11(1.40~3.17),发生脑卒中的aHR(95%CI)分别为1.45(1.11~1.71)、2.95(1.75~4.97)和2.34(1.48~3.71),发生MI的aHR(95%CI)分别为1.34(0.76~2.34)、1.17(0.62~2.19)和1.32(0.69~2.55)。分层分析结果提示只有SBP轨迹与性别和年龄的相乘交互作用有统计学意义(P交互<0.05),DBP轨迹与性别和年龄的相乘交互作用无统计学意义(P交互>0.05)。男性人群中,以SBP轨迹正常高值血压下降组作为参照组,正常高值血压稳定组、正常高值血压上升组和正常高值血压上升-下降组发生CVD发生HR(95%CI)分别为1.42(1.20~1.67)、1.93(1.53~2.42)和2.30(1.82~2.90);女性人群中,以SBP轨迹正常高值血压下降组作为参照组,正常高值血压稳定组、正常高值血压上升组和正常高值血压上升-下降组发生CVD发生HR(95%CI)分别为1.80(1.14~2.85)、2.91(1.62~5.21)和2.79(1.43~5.43)。年龄<60岁的正常高值血压人群中,以SBP轨迹正常高值血压下降组作为参照组,发生CVD的HR(95%CI)分别为1.48(1.22~1.78)、2.00(1.46~2.71)和3.01(2.24~4.06),年龄≥60岁的正常高值血压人群中,以SBP轨迹正常高值血压下降组作为参照组,发生CVD的HR(95%CI)分别为1.35(1.01~1.79)、1.94(1.40~2.69)和1.91(1.35~2.69)。

结论

在正常高值血压人群中具有CVD发病风险更高的不同血压轨迹类型。在SBP血压轨迹中,高水平且波动大的轨迹模式发生CVD的风险最高,在DBP轨迹中上升趋势且基线DBP高的轨迹模式发生CVD的风险最高。正常高值血压人群即使血压稳定在120~139/80~89 mmHg范围内且没有较大的变化波动,但依然存在发生CVD的风险。

关键词: 心血管疾病, 正常高值血压, 血压轨迹, 开滦队列研究, 群组化轨迹模型, Cox比例风险回归模型