Healthy ageing trajectories and incident cardiovascular disease in adults aged 50 years and older: a multi-cohort study Authors: Rule Du, Ruixia Niu, Yang Xu, Qian Gao, Tong Wang Published: 2026 ABSTRACT Graphical abstract Introduction As life expectancy increases and population ageing accelerates worldwide, the growing number of older adults leads to a greater burden of chronic diseases and functional decline, making healthy ageing a major public health priority [1]. In 2015, the World Health Organization (WHO) formalised the concept of healthy ageing and defined it as the process of developing and maintaining the functional ability that enables well-being in older age [2]. Building on this concept, several multidimensional measures have been developed to characterize healthy ageing, differing in their conceptual foundations and measurement approaches. Frailty-based measures primarily focus on the accumulation of health deficits [3], whereas intrinsic capacity reflects an individual’s physical and mental capacities [4]. The Ageing Trajectories of Health-Longitudinal Opportunities and Synergies (ATHLOS) score was developed in line with the WHO healthy ageing framework and incorporates multiple domains of intrinsic capacity and functional ability, providing a comprehensive assessment of healthy ageing and a common measurement framework to facilitate cross-cohort comparisons of longitudinal healthy ageing patterns [5]. In recent years, accumulating evidence has indicated that healthy ageing is closely associated with a range of adverse health outcomes, including functional limitations, disability, and increased mortality risk [[6], [7], [8]]. Among age-related diseases, cardiovascular disease (CVD) imposes a substantial global burden and represents a major public health challenge [9]. According to estimates from the Global Burden of Disease (GBD) study, approximately 626 million people were living with CVD worldwide in 2023, accounting for about 8% of the global population [10]. Understanding CVD risk from the perspective of healthy ageing may provide important insights for prevention strategies that focus on overall health and functional status rather than on individual risk factors alone. Previous studies investigating the association between healthy ageing and incident CVD have predominantly relied on health indicators assessed at a single time point, which may not adequately capture the dynamic and heterogeneous nature of healthy ageing processes [11]. Although recent studies have begun to investigate longitudinal changes or trajectories in intrinsic capacity, or related health indicators, the available evidence has mainly been derived from a single country or cohort and has often focused on selected dimensions of health [12,13]. Few studies have applied a harmonized multidimensional healthy ageing score to identify long-term trajectories and assess whether their associations with incident CVD are consistent across populations from different social and regional backgrounds. Furthermore, the potential modifying roles of socioeconomic status and lifestyle factors in this association remains unclear [14,15]. Therefore, using data from three large longitudinal ageing cohorts—Survey of Health, Ageing and Retirement in Europe (SHARE), English Longitudinal Study of Ageing (ELSA), and China Health and Retirement Longitudinal Study (CHARLS)—this study aimed to identify long-term trajectories of the ATHLOS healthy ageing score and to examine their associations with the risk of incident CVD. The inclusion of cohorts from different geographical and social settings enabled us to assess the consistency of these associations across populations. In addition, the study evaluated whether these associations varied according to socioeconomic status and lifestyle factors. Methods Study design and population SHARE is a cross-national longitudinal study of adults aged 50 years and older across multiple European countries [16]. ELSA is a nationally representative longitudinal study of adults aged 50 years and older living in England [17], whereas CHARLS is a nationally representative longitudinal study of adults aged 45 years and older in China [18]. Detailed descriptions of the study design and survey methods of these cohorts are provided in Supplementary Text S1. All cohorts obtained ethical approval and written informed consent; details are provided in the Ethics statement. In this study, wave 1 of SHARE (2004/2005), wave 1 of ELSA (2002), and wave 1 of CHARLS (2011) were defined as the baseline surveys, while wave 5 of SHARE (2013), wave 7 of ELSA (2014), and wave 3 of CHARLS (2015) were defined as the final surveys of the trajectory period. Based on repeated measurements of the ATHLOS healthy ageing score during this period, trajectories of healthy ageing were identified for each participant. Follow-up began at the end of the trajectory period. The maximum follow-up extended to wave 9 (2021) in SHARE, wave 10 (2021) in ELSA, and wave 5 (2020) in CHARLS. The process of participant selection is presented in Fig. 1.Fig. 1Study design and flowchart of participant selection. a: population selection process; b: diagram of the study design. SHARE, Survey of Health, Ageing and Retirement in Europe; ELSA, English Longitudinal Study on Ageing; CHARLS, China Health and Retirement Longitudinal Study; CVD, cardiovascular disease; ATHLOS, Ageing Trajectories of Health-Longitudinal Opportunities and Synergies.Fig. 1 Study design and flowchart of participant selection. a: population selection process; b: diagram of the study design. SHARE, Survey of Health, Ageing and Retirement in Europe; ELSA, English Longitudinal Study on Ageing; CHARLS, China Health and Retirement Longitudinal Study; CVD, cardiovascular disease; ATHLOS, Ageing Trajectories of Health-Longitudinal Opportunities and Synergies. ATHLOS healthy ageing score The ATHLOS healthy ageing score was developed based on the healthy ageing framework proposed by the WHO to assess overall health and functional status [19]. The score was constructed using an item response theory (IRT) model based on 41 items related to intrinsic capacity and functional ability, comprising 34 questionnaire-based items and seven items derived from performance-based assessments, including six cognitive tests and one objectively measured walking-speed item [5]. It was standardized to a T-score metric, with higher scores indicating better healthy ageing. Further details are provided in Supplementary Text S2. Socioeconomic status and lifestyle behaviors Composite socioeconomic status (SES) was constructed following previously established methods using baseline educational attainment and total household wealth (THW) [20,21]. Given cross-country differences in income levels, THW quartiles were calculated separately within each cohort. Education and THW scores were summed to generate an SES score ranging from 0 to 5, which was categorized into four levels: low (0), lower-middle (1–2), upper-middle (3−4), and high (5). These four levels were subsequently collapsed into two groups—low/lower-middle and upper-middle/high—for subsequent analyses because of small subgroup sizes. The healthy lifestyle score (HLS) was constructed from four baseline behavioral factors: physical activity, smoking, alcohol consumption, and social participation [22]. One point was assigned for each healthy behavior, yielding a total score ranging from 0 to 4. The score was categorized as 0−2 and 3−4 in subsequent analyses. Detailed scoring criteria are provided in Supplementary Table S1. Covariates Baseline covariates included age, sex, marital status, body mass index (BMI), hypertension, and diabetes. To improve comparability across cohorts, marital status was categorized as married or partnered and other (separated, divorced, never married, or widowed). BMI was classified as underweight, normal weight, overweight, and obesity according to WHO recommendations and applicable national or regional standards for the respective cohorts [23,24]. Outcome ascertainment and follow-up The primary outcome of this study was incident CVD during follow-up. CVD was identified based on participants’ self-reported physician diagnosis of heart disease (including angina, heart attack, congestive heart failure, and other heart problems) or stroke. Detailed definitions of CVD and the corresponding survey questions for each cohort are provided in Supplementary Text S3. Follow-up began at the final survey of the trajectory assessment period: wave 5 of SHARE, wave 7 of ELSA, and wave 3 of CHARLS. Participants were followed until the first reported CVD event, death, or their last available CVD-free survey, whichever occurred first. Follow-up time was calculated from the final trajectory survey to the corresponding event or censoring date. Statistical analyses Continuous variables were presented as medians with interquartile ranges (IQR), and categorical variables were presented as counts (percentages). Differences in baseline characteristics across healthy ageing trajectory groups were assessed using the Kruskal-Wallis test or the chi-square test. Group-based trajectory modeling (GBTM) was used to identify distinct trajectories of the ATHLOS score over time and was implemented using the lcmm package in R. Models with varying numbers of trajectory groups (2–4) and polynomial orders (linear to cubic) were compared. The optimal trajectory model was selected by jointly considering the Bayesian information criterion (BIC), Akaike information criterion (AIC), class proportions (>10%), average posterior probability (APP > 0.7), and odds of correct classification (OCC > 5) [25,26]. Missing covariates were handled using multiple imputation by chained equations (MICE). A total of 10 imputed datasets were generated, and the results were combined according to Rubin's rules [27]. Cox proportional hazards models were used to examine the association between healthy ageing trajectories and the risk of incident CVD. Two models were constructed: Model 1 was unadjusted, and Model 2 was adjusted for all baseline covariates. Kaplan-Meier curves were used to describe the cumulative incidence of CVD across different trajectory groups (Supplementary Figs. S1-S3). The proportional hazards assumption was evaluated using Schoenfeld residuals for the exposure, each covariate and the overall model [28]. No significant violation of the proportional hazards assumption was detected. Stratified analyses were conducted according to SES and HLS to evaluate potential heterogeneity in the associations between healthy ageing trajectories and incident CVD. In addition, multiplicative interaction terms were included in the fully adjusted model, and likelihood ratio tests were used to assess interactions between healthy ageing trajectories and SES or HLS. To assess the robustness of the findings, several sensitivity analyses were conducted: (i) stratified analyses by age, sex, and BMI; (ii) analyses using participants with complete covariate information; (iii) participants who developed CVD within the first year of follow-up were excluded to minimize potential reverse causation; (iv) participants with hypertension or diabetes at baseline were excluded; (v) the main analysis was repeated using a Fine-Gray competing risk model, with death treated as a competing event [29]; (vi) the analyses were repeated separately with heart disease or stroke as the outcome; (vii) E-values with corresponding 95% confidence interval limits were calculated for the estimated hazard ratios to assess the robustness of the observed associations between healthy ageing trajectory groups and incident CVD to potential unmeasured confounding; (viii) the analyses were repeated using the 26 ATHLOS items common to SHARE, ELSA, and CHARLS to assess whether the findings were robust to item selection. All statistical analyses were performed using R software (version 4.4.2), and two-sided P < 0.05 was considered statistically significant. Results Baseline characteristics of the study population According to the predefined inclusion and exclusion criteria, a total of 10,847 participants from SHARE (58.4% women; median age, 59.0 years), 5,611 from ELSA (55.7% women; median age, 57.0 years), and 7,238 from CHARLS (49.9% women; median age, 59.0 years) were included in the final analysis. Additional baseline characteristics for each cohort are presented in Table 1.Table 1Baseline characteristics of participants in the SHARE, ELSA, and CHARLS cohorts.Table 1VariablesSHARE cohort N = 10847ELSA cohort N = 5611CHARLS cohort N = 7238Follow-up years of study2004−20212002−20212011−2020Total person-years66,04030,53833,216Age, median (IQR), years59.0 (54.0−66.0)57.0 (53.0−64.0)59.0 (55.0−65.0)Sex, n (%) Male4510 (41.6)2487 (44.3)3624 (50.1) Female6337 (58.4)3124 (55.7)3614 (49.9)Marital status, n (%) Married or partnered8590 (79.2)4085 (72.8)6376 (88.1) Other marital status2257 (20.8)1526 (27.2)862 (11.9)Education, n (%) Primary4709 (43.4)1991 (35.5)6592 (91.1) Secondary3576 (33.0)1636 (29.1)580 (8.0) Tertiary2562 (23.6)1984 (35.4)66 (0.9)Total household wealth, n (%) Quartile 1 (low)2683 (24.7)1403 (25.0)1818 (25.1) Quartile 22749 (25.3)1410 (25.1)1821 (25.2) Quartile 32711 (25.0)1403 (25.0)1805 (24.9) Quartile 4 (high)2704 (25.0)1395 (24.9)1794 (24.8)BMI, n (%) Underweight93 (0.9)31 (0.6)609 (8.4) Normal weight4399 (40.5)1567 (27.9)4005 (55.3) Overweight4588 (42.3)2412 (43.0)1974 (27.3) Obesity1767 (16.3)1601 (28.5)650 (9.0)Smoking status, n (%) Never smokers5949 (54.8)2366 (42.2)4244 (58.6) Ever smokers4898 (45.2)3245 (57.8)2994 (41.4)Drinking status, n (%) Less than weekly drinking4882 (45.0)1859 (33.1)5679 (78.5) Weekly drinking or more5965 (55.0)3752 (66.9)1559 (21.5)Physical activity, n (%) Active9907 (91.3)4824 (86.0)4809 (66.4) Inactive940 (8.7)787 (14.0)2429 (33.6) Social activity, n (%)4975 (45.9)1025 (18.3)3241 (44.8) Hypertension, n (%)3001 (27.7)1451 (25.9)1531 (21.2) Diabetes, n (%)737 (6.8)227 (4.0)347 (4.8)SES, n (%) Low/Low-middle5951 (54.9)2854 (50.9)5257 (72.6) Upper-middle/High4896 (45.1)2757 (49.1)1981 (27.4)HLS, n (%) 0−2 scores5966 (55.0)4483 (79.9)3477 (48.0) 3−4 scores4881 (45.0)1128 (20.1)3761 (52.0)Abbreviation: SHARE, Survey of Health, Ageing and Retirement in Europe; ELSA, English Longitudinal Study on Ageing; CHARLS, China Health and Retirement Longitudinal Study; BMI, body mass index; SES, socioeconomic status; HLS, healthy lifestyle score; IQR, interquartile range. Table 1Baseline characteristics of participants in the SHARE, ELSA, and CHARLS cohorts.Table 1VariablesSHARE cohort N = 10847ELSA cohort N = 5611CHARLS cohort N = 7238Follow-up years of study2004−20212002−20212011−2020Total person-years66,04030,53833,216Age, median (IQR), years59.0 (54.0−66.0)57.0 (53.0−64.0)59.0 (55.0−65.0)Sex, n (%) Male4510 (41.6)2487 (44.3)3624 (50.1) Female6337 (58.4)3124 (55.7)3614 (49.9)Marital status, n (%) Married or partnered8590 (79.2)4085 (72.8)6376 (88.1) Other marital status2257 (20.8)1526 (27.2)862 (11.9)Education, n (%) Primary4709 (43.4)1991 (35.5)6592 (91.1) Secondary3576 (33.0)1636 (29.1)580 (8.0) Tertiary2562 (23.6)1984 (35.4)66 (0.9)Total household wealth, n (%) Quartile 1 (low)2683 (24.7)1403 (25.0)1818 (25.1) Quartile 22749 (25.3)1410 (25.1)1821 (25.2) Quartile 32711 (25.0)1403 (25.0)1805 (24.9) Quartile 4 (high)2704 (25.0)1395 (24.9)1794 (24.8)BMI, n (%) Underweight93 (0.9)31 (0.6)609 (8.4) Normal weight4399 (40.5)1567 (27.9)4005 (55.3) Overweight4588 (42.3)2412 (43.0)1974 (27.3) Obesity1767 (16.3)1601 (28.5)650 (9.0)Smoking status, n (%) Never smokers5949 (54.8)2366 (42.2)4244 (58.6) Ever smokers4898 (45.2)3245 (57.8)2994 (41.4)Drinking status, n (%) Less than weekly drinking4882 (45.0)1859 (33.1)5679 (78.5) Weekly drinking or more5965 (55.0)3752 (66.9)1559 (21.5)Physical activity, n (%) Active9907 (91.3)4824 (86.0)4809 (66.4) Inactive940 (8.7)787 (14.0)2429 (33.6) Social activity, n (%)4975 (45.9)1025 (18.3)3241 (44.8) Hypertension, n (%)3001 (27.7)1451 (25.9)1531 (21.2) Diabetes, n (%)737 (6.8)227 (4.0)347 (4.8)SES, n (%) Low/Low-middle5951 (54.9)2854 (50.9)5257 (72.6) Upper-middle/High4896 (45.1)2757 (49.1)1981 (27.4)HLS, n (%) 0−2 scores5966 (55.0)4483 (79.9)3477 (48.0) 3−4 scores4881 (45.0)1128 (20.1)3761 (52.0)Abbreviation: SHARE, Survey of Health, Ageing and Retirement in Europe; ELSA, English Longitudinal Study on Ageing; CHARLS, China Health and Retirement Longitudinal Study; BMI, body mass index; SES, socioeconomic status; HLS, healthy lifestyle score; IQR, interquartile range. Baseline characteristics of participants in the SHARE, ELSA, and CHARLS cohorts. Abbreviation: SHARE, Survey of Health, Ageing and Retirement in Europe; ELSA, English Longitudinal Study on Ageing; CHARLS, China Health and Retirement Longitudinal Study; BMI, body mass index; SES, socioeconomic status; HLS, healthy lifestyle score; IQR, interquartile range. Identification of healthy ageing trajectories Group-based trajectory modeling (GBTM) was applied to identify longitudinal trajectories of the ATHLOS healthy ageing score in the SHARE, ELSA, and CHARLS cohorts. In each cohort, the three-class model satisfied the predefined criteria for model fit and classification quality and was therefore retained for further analyses. Among the candidate models, the quadratic specification consistently yielded lower BIC values than the linear and cubic specifications and was thus selected as the best-fitting model (Supplementary Tables S3-S5). Based on follow-up time (years since baseline) and ATHLOS scores, three distinct healthy ageing trajectories were identified in each cohort (Fig. 2). All trajectories showed a gradual declining trend over time and were classified as low, middle, and high trajectory groups.Fig. 2Trajectories of the ATHLOS healthy ageing score in the SHARE, ELSA, and CHARLS cohorts. SHARE, Survey of Health, Ageing and Retirement in Europe; ELSA, English Longitudinal Study on Ageing; CHARLS, China Health and Retirement Longitudinal Study.Fig. 2 Trajectories of the ATHLOS healthy ageing score in the SHARE, ELSA, and CHARLS cohorts. SHARE, Survey of Health, Ageing and Retirement in Europe; ELSA, English Longitudinal Study on Ageing; CHARLS, China Health and Retirement Longitudinal Study. Baseline characteristics of trajectory groups Based on the trajectories of the healthy ageing score, participants in all three cohorts were classified into low, middle, and high trajectory groups. The baseline characteristics of each trajectory group are presented in Supplementary Tables S6-S8. Compared with the high trajectory group, participants in the low trajectory group were older, more likely to be female, and had lower levels of education and household wealth. They also exhibited lower levels of physical activity and higher prevalence of hypertension and diabetes. These patterns were generally consistent across the three cohorts. Association between healthy ageing trajectories and CVD During follow-up, 1,839, 549, and 1,494 incident CVD events were observed in the SHARE, ELSA, and CHARLS cohorts, respectively. Compared with the low healthy ageing trajectory group, both the middle and high trajectory groups were associated with lower risks of CVD in the unadjusted model. In the fully adjusted model, the middle trajectory group was associated with lower risks of incident CVD (SHARE, HR: 0.89, 95% CI: 0.77–1.02; ELSA, HR: 0.66, 95% CI: 0.53−0.82; CHARLS, HR: 0.67, 95% CI: 0.58−0.76), whereas the high trajectory group was associated with further reductions in CVD risk (SHARE, HR: 0.61, 95% CI: 0.52−0.71; ELSA, HR: 0.36, 95% CI: 0.28−0.46; CHARLS, HR: 0.39, 95% CI: 0.33−0.45) (Fig. 3). Overall, a graded decrease in CVD risk was observed across increasing levels of healthy ageing trajectories.Fig. 3Hazard ratios for incident CVD across healthy ageing trajectory groups in the SHARE, ELSA, and CHARLS cohorts. SHARE, Survey of Health, Ageing and Retirement in Europe; ELSA, English Longitudinal Study on Ageing; CHARLS, China Health and Retirement Longitudinal Study, HR, hazard ratio; CI, confidence interval.Fig. 3 Hazard ratios for incident CVD across healthy ageing trajectory groups in the SHARE, ELSA, and CHARLS cohorts. SHARE, Survey of Health, Ageing and Retirement in Europe; ELSA, English Longitudinal Study on Ageing; CHARLS, China Health and Retirement Longitudinal Study, HR, hazard ratio; CI, confidence interval. Stratified analyses by SES and HLS In stratified analyses by SES and HLS, higher healthy ageing trajectories were generally associated with lower risks of incident CVD across strata in the three cohorts (Fig. 4). No statistically significant interaction between healthy ageing trajectories and SES was observed in any cohort. For HLS, a significant interaction was observed only in the SHARE cohort (P for interaction = 0.027), whereas no significant interaction was observed in ELSA or CHARLS. Detailed estimates are provided in Supplementary Table S9.Fig. 4Hazard ratios for incident CVD across healthy ageing trajectory groups stratified by SES and HLS. Panels (a–c) present results stratified by SES in SHARE, ELSA, CHARLS, respectively. Panels (d–f) present results stratified by HLS in SHARE, ELSA, CHARLS, respectively. SHARE, Survey of Health, Ageing and Retirement in Europe; ELSA, English Longitudinal Study on Ageing; CHARLS, China Health and Retirement Longitudinal Study; SES, socioeconomic status; HLS, healthy lifestyle score.Fig. 4 Hazard ratios for incident CVD across healthy ageing trajectory groups stratified by SES and HLS. Panels (a–c) present results stratified by SES in SHARE, ELSA, CHARLS, respectively. Panels (d–f) present results stratified by HLS in SHARE, ELSA, CHARLS, respectively. SHARE, Survey of Health, Ageing and Retirement in Europe; ELSA, English Longitudinal Study on Ageing; CHARLS, China Health and Retirement Longitudinal Study; SES, socioeconomic status; HLS, healthy lifestyle score. Additional subgroup and sensitivity analyses The associations remained consistent across multiple sensitivity and subgroup analyses, including stratified analyses by age, sex, and BMI (Figs. S4–S6); complete-case analysis (Table S10); exclusion of participants who developed CVD within the first year of follow-up (Table S11); exclusion of participants with baseline hypertension or diabetes (Table S12); Fine-Gray competing risk models accounting for death (Table S13); separate analyses for heart disease and stroke (Figs. S7-S8); and analyses repeated using the 26 ATHLOS items common to SHARE, ELSA, and CHARLS (Fig. S9). E-value analyses further indicated the robustness of the findings to potential unmeasured confounding, as an unmeasured confounder would need to be strongly associated with both healthy ageing trajectories and incident CVD to fully explain the observed associations (Table S14). Discussion This study, based on three large longitudinal ageing cohorts—SHARE, ELSA, and CHARLS—examined the association between healthy ageing trajectories and the risk of incident cardiovascular disease (CVD). The findings indicated that higher healthy ageing trajectories were consistently associated with lower risks of CVD across all three cohorts. Stratified analyses further showed that these associations were generally consistent across populations with different SES and HLS, with no statistically significant interactions observed. To our knowledge, this study is the first to examine the association between healthy ageing trajectories and the risk of incident CVD. The ATHLOS healthy ageing score is based on an integrated framework of intrinsic capacity (IC) and functional ability and provides a comprehensive assessment of healthy ageing at the individual level [5]. In this study, the ATHLOS score showed an overall declining trend with increasing age, which is consistent with previous studies in middle-aged and older populations. For example, a cross-national study of individuals aged 50 years and older from 15 countries in the SHARE cohort [30], as well as a study from the Yale Ageing Project involving participants aged 70 years and older [31], reported similar age-related declines in overall health. In addition, three distinct healthy ageing trajectories (low, middle, and high) were identified across all three cohorts. In the SHARE and ELSA cohorts, the high trajectory group accounted for the largest proportion of participants, whereas in the CHARLS cohort, the middle trajectory group was the most prevalent. Similar trajectory patterns have also been reported in studies of older adults in China [12]. These differences may be related to variations in socioeconomic development, access to healthcare, educational attainment, and life-course health exposures, such as early-life adversity, across countries. Previous cross-national studies have shown that older populations in high-income countries generally exhibit better functional ability and overall health, which may partly explain the higher proportion of individuals in the high trajectory group observed in the European cohorts [32,33]. This study further demonstrated that higher healthy ageing trajectories were significantly associated with lower risks of incident CVD, suggesting that overall health status and its long-term changes may play an important role in cardiovascular health among middle-aged and older adults. The present findings are consistent with previous studies examining multidimensional indicators of health and functional status in relation to cardiovascular outcomes. For example, prior studies have shown that older adults with higher frailty index (FI) or lower IC tend to have higher risks of CVD and mortality, highlighting the importance of overall health status in the development of chronic diseases [6,34,35]. A recent review further suggests that individuals with higher IC levels generally have better functional reserves and physiological resilience, which may enhance their ability to cope with disease or physiological stress and thereby reduce the risk of chronic diseases [36]. From a biological perspective, better healthy ageing status is often associated with lower levels of inflammatory markers, such as IL-6, C-reactive protein (CRP), and tumor necrosis factor-α (TNF-α), which may help mitigate endothelial dysfunction and slow the progression of atherosclerosis, thereby reducing the risk of CVD [37]. Further analyses indicated that no statistically significant modifying effects of SES or HLS on the association between healthy ageing and incident CVD were observed. This finding may be explained by the fact that the healthy ageing score, as a multidimensional composite measure, partly captures the cumulative effects of socioeconomic conditions and long-term lifestyle exposures on health, thereby limiting the additional modifying effects of SES and HLS in this association [33]. The findings of this study contribute to understanding the relationship between healthy ageing and the risk of chronic diseases. From a life-course perspective, health status in later life is not static but follows distinct long-term trajectories, and identifying these patterns may provide a more comprehensive understanding of health decline or maintenance. This study further indicated that healthy ageing not only reflects overall health status in older age but is also closely associated with subsequent risks of chronic diseases. Therefore, analyzing healthy ageing trajectories as a composite health indicator may help identify high-risk populations and inform targeted strategies for health promotion and disease prevention in older adults. This study has several strengths. First, it was based on three large longitudinal ageing cohorts—SHARE, ELSA, and CHARLS—with substantial sample sizes and diverse national backgrounds, which enhancing the robustness and external generalizability of the findings. Second, group-based trajectory modeling was applied to identify long-term patterns of healthy ageing, enabling assessment of individual health status from a dynamic perspective and providing a novel approach to understanding health changes over time. In addition, multiple sensitivity analyses were conducted, and the results were consistent with the main findings, further supporting the robustness of the conclusions. Several limitations should also be considered. First, consistent with previous studies, CVD outcomes were primarily based on self-reported physician diagnoses, which may be subject to misclassification [38,39]. Second, although a range of potential confounders was adjusted for, residual confounding cannot be completely ruled out. Finally, interaction results for lifestyle factors were not consistent across cohorts, suggesting that these findings require further validation in future studies. Conclusion Based on three longitudinal ageing cohorts—SHARE, ELSA, and CHARLS—this study demonstrated that higher healthy ageing trajectories were significantly reduced the risks of incident cardiovascular disease (CVD), and this association remained consistent across populations with different socioeconomic status and lifestyle levels. These findings highlight the importance of long-term, multidimensional health trajectories in shaping CVD risk and support the integration of dynamic measures healthy ageing into risk stratification and early prevention strategies. CRediT authorship contribution statement Rule Du: Writing-original draft, Writing-review & editing, Methodology, Visualization, Data Curation. Ruixia Niu: Writing-original draft, Writing-review & editing, Methodology, Visualization, Data Curation. Yang Xu: Visualization, Data Curation. Qian Gao: Writing-review & editing, Supervision, Funding acquisition. Tong Wang: Writing-review & editing, Supervision, Funding acquisition. All authors read and approved the final manuscript. Consent for publication Not applicable. Ethics approval and consent to participate This study was a secondary analysis of publicly available datasets; therefore, additional ethical review and approval were not required. All original studies were approved by their respective institutional review boards, and participants provided written informed consent. SHARE was approved by the Ethics Review Committee of the University of Mannheim. ELSA was approved by the London Multi-Centre Research Ethics Committee (MREC/01/2/91; MREC/04/2/006; 05/MRE02/63). CHARLS was approved by the Ethics Review Committee of Peking University (IRB00001052-11015). Declaration of Generative AI and AI-assisted technologies in the writing process During the preparation of this work, the authors used ChatGPT-5.6 solely for grammar correction. After using this tool, the authors reviewed and edited the content as necessary and take full responsibility for the content of the publication. Funding sources This study was supported by the 10.13039/501100001809National Natural Science Foundation of China (Grant numbers: 82373692, 82574210, 82073674, and 82204163), and Shanxi Province Higher Education “Billion Project” Science and Technology Guidance Project. Data availability statement The datasets used in this study are available upon request by submitting a data user agreement to the respective data teams via the websites of the Survey of Health, Ageing and Retirement in Europe (https://releases.sharedataportal.eu/), the English Longitudinal Study of Ageing (https://ukdataservice.ac.uk/), and the China Health and Retirement Longitudinal Study (https://charls.pku.edu.cn/). Declaration of competing interest The authors declare no competing interests.