Abstract
Background
Cardiovascular disease (CVD) remains a leading global cause of mortality, with obesity being a major modifiable risk factor. The recent European Association for the Study of Obesity (EASO) framework diagnosed obesity by integrating central adiposity and specific complications, advancing beyond body mass index (BMI) alone. However, the association between longitudinal obesity trajectories defined by this framework and CVD risk is unknown.
Methods
We utilized longitudinal data from the China Health and Retirement Longitudinal Study and the English Longitudinal Study of Ageing. Participants were considered to have obesity if they had a BMI of 30 kg/m2 or higher, or a BMI between 25 and < 30 kg/m2 with waist‑to‑height ratio ≥ 0.5 plus at least one of hypertension, diabetes, arthritis, renal disease, chronic obstructive pulmonary disease, or depression. First, we explored the association between baseline obesity and CVD risk. Second, based on three waves, participants were categorized into five trajectories including no obesity, decreasing obesity, increasing obesity, consistent obesity, and fluctuating obesity. Cox proportional hazards models were used to examine the associations between different obesity trajectories and CVD.
Results
A total of 11,858 participants were included in the study. Compared with those without baseline obesity, participants with baseline obesity had a 36% (hazard ratio [HR]: 1.36, 95% confidence interval [CI]: 1.26–1.47, p < 0.001) increased risk of CVD incidence. Besides, based on 7,154 enrolled participants, we found that compared to the no obesity group, participants with a consistent and increasing obesity trajectory had a 50% (HR: 1.50, 95% CI: 1.30–1.72, p < 0.001) and 34% (HR: 1.34, 95% CI: 1.11–1.62, p = 0.003) higher risk of CVD. In contrast, decreasing and fluctuating obesity trajectories were not significantly associated with CVD risk.
Conclusion
Based on the novel EASO definition of obesity, our study confirmed that baseline obesity was associated with increased CVD risk. Beyond this, dynamic trajectory analyses revealed that both consistent and increasing obesity trajectories were significantly associated with elevated CVD risk. These findings suggest that in CVD prevention, clinical attention should focus on not only the presence of obesity, but also its dynamic progression over time.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12889-026-27213-7.
Keywords: Cardiovascular disease, Obesity trajectory, China Health and Retirement Longitudinal Study, English Longitudinal Study of Ageing
Introduction
Although there have been substantial advancements in scientific studies and pharmacological treatments, cardiovascular disease (CVD) remains the foremost cause of death around the world, with the number of deaths climbing from 12.1 million in 1990 to 20.5 million in 2025, and projected to reach 35.6 million by 2050 [1, 2]. Adjustable risk factors for CVD [3], including hyperlipidemia, hypertension, type 2 diabetes mellitus, obesity, and smoking, account for 57.2% of the global CVD burden in women and 52.6% in men [4]. Among these modifiable factors, obesity represents a particularly pervasive and potent driver of cardiovascular risk. Obesity constitutes a critical risk factor for a range of CVD, including atherosclerosis, coronary heart disease, and heart failure [5]. It is estimated that approximately 880 million adults globally, equating to 16% of the world’s population, meet the criterion for obesity using the threshold of a body mass index (BMI) of 30 kg/m2 or higher. Consequently, obesity is contributing to a growing and substantial portion of the worldwide CVD burden. This underscores obesity management as a critical public health priority for CVD prevention.
Despite its significant role, the clinical diagnosis of obesity has long relied predominantly on body mass index (BMI) cut-off values, a simplistic metric that fails to capture the distribution and functionality of adipose tissue, which are key determinants of health risk [6, 7]. This anthropometric-only approach overlooks the pathophysiological essence of obesity as a chronic disease. To address this critical limitation, the European Association for the Study of Obesity (EASO) recently proposed a transformative diagnostic framework. This new definition refines the diagnosis of obesity by integrating an anthropometric component, which uses waist-to-height ratio to assess central adiposity, with a clinical component that requires the presence of medical, functional, or psychological complications, including hypertension, CVD, diabetes, arthritis, renal disease, chronic obstructive pulmonary disease, and depression [8]. These complications were selected by the EASO consensus panel as the most common and clinically significant obesity-related conditions, each with well-established pathophysiological links to excess adiposity. This integrated approach enables a more accurate identification of individuals whose excess adiposity constitutes a genuine disease state.
Building upon this enhanced diagnostic framework, a critical next step is to move from static assessment to dynamic evaluation. Obesity is a chronic and progressive condition [9]. Therefore, understanding an individual’s obesity trajectory, characterized by patterns of stability, increase, decrease, or fluctuation in obesity status over time, is vital. Such longitudinal assessment may reveal distinct risk profiles that a single measurement obscures, offering deeper insights for personalized cardiovascular risk prediction and long-term management strategies. However, the relationship between obesity trajectories defined by this new EASO framework and incident CVD remains unexplored. To address this critical gap, the present study leveraged two large prospective cohorts, namely the China Health and Retirement Longitudinal Study (CHARLS) and the English Longitudinal Study of Ageing (ELSA). We aim to characterize distinct obesity trajectories based on the EASO criteria across multiple survey waves and to investigate their specific associations with the prevalence of CVD.
Methods
Study design and population
The data used in this study were derived from CHARLS (https://charls.pku.edu.cn/en/) and ELSA (https://www.elsa-project.ac.uk/). The CHARLS, which began in June 2011, involves surveying a nationally representative sample of Chinese people aged 45 and older, with follow-ups occurring every two to three years, and is approved by Peking University’s Ethical Review Committee [10]. The ELSA is a panel study initiated in 1998, tracking representative groups of people aged 50 and above living in England, with follow-ups every two years to gather data on the aging population. CHARLS and ELSA were selected as they are among the few longitudinal aging cohorts that provide complete and harmonized data across multiple waves on all components required to operationalize the novel EASO diagnostic framework, including anthropometric measures and a comprehensive set of clinical comorbidities. Their parallel design, representing distinct geographic regions and populations, enhances the generalizability of our findings. Ethical approval and consent were obtained for all waves and components [11].
A total of 26,644 participants from CHARLS and ELSA were initially considered for this study. This research utilized wave 1 (2011) of CHARLS and wave 2 (2004) of ELSA as the baseline. The second survey included wave 2 (2013) of CHARLS and wave 4 (2008) of ELSA. The third survey comprised wave 3 (2015) of CHARLS and wave 6 (2012) of ELSA. Follow-up assessments were carried out until wave 5 (2020) for CHARLS and wave 9 (2018) for ELSA. We excluded participants with baseline CVD or missing CVD information (n = 4,779), lacked baseline obesity data (n = 6,273), or were not followed up at the end of the survey (n = 3,734). In our study, “not followed up” was defined as absence of any interview in waves following the exposure period. These participants were excluded because they contributed no person-time to the time-to-event analysis. A total of 11,858 participants were included in this study to analyze the association between baseline obesity and the risk of CVD (Fig. 1). We then excluded participants with CVD or missing CVD information during the exposure period (waves 1–3 for CHARLS and waves 2–6 for ELSA) (n = 11,700), without information on obesity during the exposure period (n = 6,820), or loss to follow-up at the end survey (n = 970). Ultimately, 7,154 participants were included in the analysis examining the link between obesity trajectory and CVD risk (Fig. 2).
Fig. 1.
Flow chart of inclusion and exclusion of the participants for the association between baseline obesity and follow-up CVD. Abbreviation: CHARLS, China Health and Retirement Longitudinal Study; CVD, cardiovascular disease; ELSA, English Longitudinal Study of Ageing
Fig. 2.
Flow chart of inclusion and exclusion of the participants for the association between obesity trajectories and follow-up CVD. Abbreviation: CHARLS, China Health and Retirement Longitudinal Study; CVD, cardiovascular disease; ELSA, English Longitudinal Study of Ageing
Definition of the new framework of obesity
In this study, obesity was defined according to the EASO new framework definition [8]. Participants were considered with obesity if they had a BMI ≥ 30 kg/m2, or a BMI between 25 and < 30 kg/m² along with a waist‑to‑height ratio (WHtR) ≥ 0.5, and the presence of one or more medical, functional, or psychological comorbidities, including hypertension, CVD, diabetes, arthritis, renal disease, chronic obstructive pulmonary disease, and depression. These comorbidities represent the clinical manifestations of obesity as a chronic disease and are integral components of the diagnostic criteria. Since the study focused on exploring the link between obesity and CVD risk, the outcome variable “CVD” was excluded from the obesity definition.
In addition, there are some differences in the definition of obesity between the two cohorts. For CHARLS participants, the corresponding BMI cut‑offs are adjusted to ≥ 27.5 kg/m² and 23 to < 27.5 kg/m², respectively, as a lower anthropometric cut point is advised given their predisposition to developing metabolic abnormalities at lower levels of adiposity [12]. For participants in ELSA, given the absence of renal disease-related information, the diagnostic criteria for obesity excluded renal disease.
BMI was determined by dividing weight in kilograms by the square of height in meters. WHtR was calculated by dividing waist circumference in centimeters by height in centimeters. A diagnosis of hypertension was made if participants met any of the following criteria: blood pressure ≥ 130/80 mmHg, a self-reported history of hypertension, or current antihypertensive medication use [13]. Similarly, diabetes was diagnosed based on a fasting blood glucose ≥ 126 mg/dL, an HbA1c ≥ 6.5%, a prior self-reported diagnosis, or current hypoglycemic treatment [14]. Renal disease was defined based on self-report or a reduced estimated glomerular filtration rate (< 60 mL/min/1.73 m²), while chronic obstructive pulmonary disease and arthritis were defined by self-reported diagnosis only. In addition, depressive symptoms are evaluated with the Center for Epidemiologic Studies Depression (CES-D), a scale recognized for its reliable and valid measurements [15]. The CES-D score ranges varied between CHARLS (0–30) and ELSA (0–8), with higher scores indicating more severe symptoms. In CHARLS, a score of 10 or more (CES-D ≥ 10) was utilized to signify elevated depressive symptoms, whereas in ELSA, a score of 3 or above (CES-D ≥ 3) was identified as indicating clinically significant depressive symptoms, consistent with prior studies [16, 17].
Assessment of obesity trajectories
To develop a detailed obesity construct that fully represents obesity status, we combined baseline obesity data with longitudinal changes. Participants’ obesity variations from CHALRS waves 1–3 and ELSA waves 2–6 were analyzed to create five obesity trajectories. The term “no obesity” was characterized as the complete absence of obesity across all assessed time points. The term “decreasing obesity” was defined by two patterns: (1) the presence of obesity at the initial time point followed by an absence of obesity at the subsequent two time points, and (2) the presence of obesity at the first two time points with an absence of obesity at the final time point. Conversely, “increasing obesity” was defined by: (1) an absence of obesity at the initial time point with the presence of obesity at the following two time points, and (2) an absence of obesity at the first two time points with the presence of obesity at the final time point. The term “consistent obesity” referred to the continuous obesity across the entire assessment period. Lastly, “fluctuating obesity” was used to describe obesity trajectories that do not align with the aforementioned classifications (Table S1). The group with a no obesity trajectory was considered the reference group. For the analysis, we employed a rule-based trajectory classification approach following established methods in longitudinal epidemiological research [18–20]. This method prioritizes clinical interpretability and ensures consistent application across the two cohorts.
Determination of CVD
The outcome event in this study was a self-reported diagnosis of CVD, specifically if participants were informed by a doctor about having heart disease or stroke [21, 22]. Participants were asked in each wave, “Have you been told by a doctor that you have been diagnosed with a heart disease, including angina, heart attack, congestive heart failure, and other heart problems?” and “Have you been told by a doctor that you have been diagnosed with a stroke?” Those who indicated a diagnosis of heart disease or stroke were identified as having CVD. It is important to note that CVD status was assessed at multiple time points: baseline CVD was used for exclusion in baseline obesity analyses, CVD during the exposure period was used for exclusion in obesity trajectory analyses, and incident CVD occurring after the respective exclusion periods was treated as the outcome event in Cox regression models.
Covariates
The following covariates were included: age, sex, marital status, education attainment, smoking status, and drinking status. For consistency among CHARLS and ELSA, marital status was divided into two groups: married or partnered and other marital status (separated, divorced, unmarried, or widowed). Education attainment was classified into three levels: below high school, high school, and college or above. Smoking status was grouped into current smokers and those who have either quit or never smoked. Similarly, drinking status was categorized into current drinkers and those who were former or never drinkers.
Statistical analysis
The Kolmogorov–Smirnov test was employed to evaluate the normality of distributions for continuous variables. Besides, we evaluated skewness and kurtosis. If skewness was above + 1 or below − 1, and kurtosis exceeded + 1, the data distribution typically did not follow a normal distribution. Continuous variables were described according to their distributional characteristics: variables with normal distributions were presented as means and standard deviations, while those with skewed distributions were reported as medians and interquartile ranges. For group comparisons, the choice of statistical tests was based on the number of groups and the distribution of the data. For comparisons between two groups (Table 1), continuous variables were analyzed using either the independent samples Student’s t-test (for normally distributed data) or the Mann-Whitney U-test (for non-normally distributed data). For comparisons across multiple groups (Table 2), analysis of variance or the Kruskal-Wallis test was applied, as appropriate. Categorical variables were presented as percentages and compared between groups using the Chi-square test.
Table 2.
Baseline characteristics of participants and incident CVD during follow-up for obesity trajectory analyses
| Characteristic | Overall (n = 7154) | No obesity (n = 2841) | Decreasing obesity (n = 371) | Fluctuating obesity (n = 432) | Increasing obesity (n = 928) | Consistent obesity (n = 2582) | p value |
|---|---|---|---|---|---|---|---|
| Age, years | 58.73(7.94) | 58.76(7.97) | 58.98(8.44) | 58.62(8.19) | 58.23(8.02) | 58.85(7.74) | 0.320 |
| Body mass index, kg/m2 | 25.08(4.63) | 21.68(2.66) | 26.08(3.97) | 24.78(2.95) | 24.52(2.42) | 28.92(4.25) | < 0.001 |
| Waist-to-height ratio | 0.55(0.08) | 0.50(0.06) | 0.57(0.04) | 0.55(0.06) | 0.53(0.07) | 0.60(0.06) | < 0.001 |
| Sex | < 0.001 | ||||||
| Female | 3959(55.34) | 1427(50.23) | 217(58.49) | 256(59.26) | 522(56.25) | 1537(59.53) | |
| Male | 3195(44.66) | 1414(49.77) | 154(41.51) | 176(40.74) | 406(43.75) | 1045(40.47) | |
| Marital status | 0.182 | ||||||
| Married or partnered | 5996(83.81) | 2404(84.62) | 312(84.10) | 370(85.65) | 757(81.57) | 2153(83.38) | |
| Other marital status | 1158(16.19) | 437(15.38) | 59(15.90) | 62(14.35) | 171(18.43) | 429(16.62) | |
| Education attainment | < 0.001 | ||||||
| Below high school | 5059(70.72) | 2115(74.45) | 276(74.39) | 316(73.15) | 629(67.78) | 1723(66.73) | |
| High school | 1165(16.28) | 401(14.11) | 54(14.56) | 58(13.43) | 185(19.94) | 467(18.09) | |
| College or above | 930(13.00) | 325(11.44) | 41(11.05) | 58(13.43) | 114(12.28) | 392(15.18) | |
| Smoking status | < 0.001 | ||||||
| Non-current smoker | 5409(75.61) | 1907(67.12) | 281(75.74) | 334(77.31) | 717(77.26) | 2170(84.04) | |
| Current smoker | 1745(24.39) | 934(32.88) | 90(24.26) | 98(22.69) | 211(22.74) | 412(15.96) | |
| Alcohol status | 0.231 | ||||||
| Non-current drinker | 3451(48.24) | 1377(48.47) | 188(50.67) | 227(52.55) | 443(47.74) | 1216(47.10) | |
| Current drinker | 3703(51.76) | 1464(51.53) | 183(49.33) | 205(47.45) | 485(52.26) | 1366(52.90) | |
| Hypertension | < 0.001 | ||||||
| No | 4534(63.38) | 2300(80.96) | 190(51.21) | 280(64.81) | 734(79.09) | 1030(39.89) | |
| Yes | 2620(36.62) | 541(19.04) | 181(48.79) | 152(35.19) | 194(20.91) | 1552(60.11) | |
| Diabetes | < 0.001 | ||||||
| No | 6548(91.53) | 2706(95.25) | 314(84.64) | 392(90.74) | 889(95.80) | 2247(87.03) | |
| Yes | 606(8.47) | 135(4.75) | 57(15.36) | 40(9.26) | 39(4.20) | 335(12.97) | |
| Arthritis | < 0.001 | ||||||
| No | 4969(69.46) | 2167(76.28) | 252(67.92) | 325(75.23) | 731(78.77) | 1494(57.86) | |
| Yes | 2185(30.54) | 674(23.72) | 119(32.08) | 107(24.77) | 197(21.23) | 1088(42.14) | |
| COPD | < 0.010 | ||||||
| No | 6720(93.93) | 2673(94.09) | 337(90.84) | 413(95.60) | 889(95.80) | 2408(93.26) | |
| Yes | 434(6.07) | 168(5.91) | 34(9.16) | 19(4.40) | 39(4.20) | 174(6.74) | |
| Depression | < 0.001 | ||||||
| No | 5392(75.37) | 2100(73.92) | 232(62.53) | 327(75.69) | 771(83.08) | 1962(75.99) | |
| Yes | 1762(24.63) | 741(26.08) | 139(37.47) | 105(24.31) | 157(16.92) | 620(24.01) | |
| CVD | < 0.001 | ||||||
| No | 6025(84.22) | 2506(88.21) | 309(83.29) | 371(85.88) | 769(82.87) | 2070(80.17) | |
| Yes | 1129(15.78) | 335(11.79) | 62(16.71) | 61(14.12) | 159(17.13) | 512(19.83) |
Values are mean (standard deviation) for continuous variables or numbers (%) for categorical variables
“CVD = Yes” indicates participants who developed incident CVD during follow-up after the exposure period (through CHARLS wave 5, 2020; ELSA wave 9, 2018). Participants with CVD during the exposure period (CHARLS waves 1–3, 2011–2015; ELSA waves 2–6, 2004–2012) were excluded from the trajectory analysis
Abbreviations: COPD chronic obstructive pulmonary disease, CVD cardiovascular disease
Table 1.
Baseline characteristics of study participants and incident CVD during follow-up for baseline obesity analyses
| Characteristic | Total (N = 11858) |
Non-obesity (N = 6721) |
Obesity (N = 5137) |
p value |
|---|---|---|---|---|
| Age, years | 58.48(8.33) | 58.38(8.43) | 58.61(8.20) | 0.136 |
| Body mass index, kg/m2 | 24.68(4.50) | 22.14(2.75) | 28.00(4.16) | < 0.001 |
| Waist-to-height ratio | 0.54(0.08) | 0.50(0.07) | 0.59(0.06) | < 0.001 |
| Sex | < 0.001 | |||
| Female | 6496(54.78) | 3403(50.63) | 3093(60.21) | |
| Male | 5362(45.22) | 3318(49.37) | 2044(39.79) | |
| Marital status | 0.364 | |||
| Married or partnered | 10,108(85.24) | 5747(85.51) | 4361(84.89) | |
| Other marital status | 1750(14.76) | 974(14.49) | 776(15.11) | |
| Education attainment | < 0.001 | |||
| Below high school | 8957(75.54) | 5187(77.18) | 3770(73.39) | |
| High school | 1758(14.82) | 949(14.12) | 809(15.75) | |
| College or above | 1143(9.64) | 585(8.70) | 558(10.86) | |
| Smoking status | < 0.001 | |||
| Non-current smoker | 8743(73.73) | 4569(67.98) | 4174(81.25) | |
| Current smoker | 3115(26.27) | 2152(32.02) | 963(18.75) | |
| Alcohol status | 0.212 | |||
| Non-current drinker | 6297(53.10) | 3535(52.60) | 2762(53.77) | |
| Current drinker | 5561(46.90) | 3186(47.40) | 2375(46.23) | |
| Hypertension | < 0.001 | |||
| No | 7421(62.58) | 5261(78.28) | 2160(42.05) | |
| Yes | 4437(37.42) | 1460(21.72) | 2977(57.95) | |
| Diabetes | < 0.001 | |||
| No | 10,783(90.93) | 6371(94.79) | 4412(85.89) | |
| Yes | 1075(9.07) | 350(5.21) | 725(14.11) | |
| Arthritis | < 0.001 | |||
| No | 8050(67.89) | 5014(74.60) | 3036(59.10) | |
| Yes | 3808(32.11) | 1707(25.40) | 2101(40.90) | |
| COPD | 0.109 | |||
| No | 11,053(93.21) | 6287(93.54) | 4766(92.78) | |
| Yes | 805(6.79) | 434(6.46) | 371(7.22) | |
| Depression | < 0.001 | |||
| No | 8534(71.97) | 4941(73.52) | 3593(69.94) | |
| Yes | 3324(28.03) | 1780(26.48) | 1544(30.06) | |
| CVD | < 0.001 | |||
| No | 9037(76.21) | 5392(80.23) | 3645(70.96) | |
| Yes | 2821(23.79) | 1329(19.77) | 1492(29.04) |
Values are mean (standard deviation) for continuous variables or numbers (%) for categorical variables
“CVD = Yes” indicates participants who developed incident CVD during follow-up (through CHARLS wave 5, 2020; ELSA wave 9, 2018) after being free of CVD at baseline. Participants with baseline CVD were excluded from the analysis
Abbreviations: COPD chronic obstructive pulmonary disease, CVD cardiovascular disease
The Cox proportional hazards model was used to calculate the hazard ratio (HR) and 95% confidence interval (CI) to assess the association of baseline obesity and obesity trajectory with CVD risk. Three models were fitted: Model 1 was unadjusted; Model 2 was adjusted for demographic confounders, including age, sex, marital status, education attainment; Model 3 was further adjusted for lifestyle factors including smoking and drinking status. Cox proportional hazards regression assumptions were tested using Schoenfeld residuals.
Subgroup analysis was performed using the following covariates: age, sex, marital status, education attainment, smoking status, and drinking status. For the sensitivity analysis, first, the main analyses were conducted with heart disease and stroke considered as separate outcomes. Second, to examine potential regional differences, we performed sensitivity analyses stratified by cohort (CHARLS vs. ELSA), repeating the main analyses separately for each database. Third, to empirically assess whether residual confounding by comorbidities such as hypertension or diabetes might influence our results, we performed a sensitivity analysis restricted to participants classified as obese by BMI ≥ 30 kg/m² only (i.e., those whose obesity status does not depend on the presence of comorbidities). Fourth, to account for death during follow-up as a competing risk, a Fine–Gray competing risks model was constructed.
To quantify the added value of the EASO framework over the traditional BMI-based definition, we performed several analyses. First, we created a confusion matrix cross-tabulating obesity classification by BMI-based versus EASO criteria. Second, we compared the predictive performance for CVD risk through the use of receiver operating characteristic curves and continuous net reclassification improvement (NRI), with BMI-based obesity as the reference.
Statistical significance was recognized with a two-sided p-value of less than 0.050, and R 4.5.0 was used for conducting all statistical analyses in this study.
Results
Baseline characteristics
A total of 11,858 participants were included in this study. The mean age of the participants was 58.48 years, with a 54.78% female predominance. Table 1 displayed the sample characteristics based on baseline obesity categories, with 56.68% of participants categorized as non-obesity and 43.32% as obesity. Compared with non-obesity participants, obesity participants were more likely to be females, non-current smokers, have more highly educated people (all p < 0.001). Additionally, they tended to have higher prevalence rates of hypertension, diabetes, arthritis, depression, and CVD (all p < 0.001), with the exception of COPD which showed no significant difference between groups (p = 0.109). Besides, as shown in Tables 2, 7 and 154 participants were enrolled in the obesity trajectory analysis. Among the included participants, 3,959 (55.34%) were females, and the mean age was 58.73 years. Five distinct obesity trajectory groups were identified, described as no obesity (n = 2,841, 39.71%), decreasing obesity (n = 371, 5.19%), fluctuating obesity (n = 432, 6.04%), increasing obesity (n = 928, 12.97%), and consistent obesity (n = 2,582, 36.09%). Individuals in the consistent obesity trajectory group were more likely to be females, non-current smokers, have more highly educated people, have elevated rates of hypertension, diabetes, arthritis, COPD, depression, and CVD, compared to those in other obesity trajectory groups (all p < 0.010). These p values indicated significant differences in baseline characteristics across baseline obesity groups and obesity trajectory groups. These descriptive differences highlighted the distinct profiles of individuals following different obesity trajectories and underscored the importance of multivariable adjustment, which we performed in subsequent analyses, to isolate the independent association between each trajectory and CVD risk.
Comparison of included and excluded participants revealed that compared to the included participants, the excluded participants were older, more likely to be male, more likely to be unmarried, and had lower educational attainment (all p < 0.050) (Table S2 and Table S3). We further compared the baseline characteristics of participants between the two cohorts (CHARLS and ELSA). As shown in Table S4 and Table S5, CHARLS participants were slightly younger, had lower BMI and WHtR, higher prevalence of married status and lower educational levels, higher smoking rates, higher prevalence of diabetes, arthritis, COPD, and depression compared to ELSA participants (all p < 0.050). These differences reflected the distinct population characteristics and healthcare contexts of the two countries.
Baseline obesity and its association with CVD
Table 3 presented the association between baseline obesity and the risk of incident CVD. Following adjusting age, sex, marital status, education attainment, smoking status, and drinking status, compare to participants without baseline obesity, the adjusted HR for CVD was 1.36 (95% CI: 1.26–1.47; p < 0.001) for those with baseline obesity. The p value represented the probability of observing such an association by chance if the null hypothesis of no association (HR = 1) were true, with p < 0.050 indicating statistical significance.
Table 3.
Association between baseline obesity and risk of CVD
| Group | Model 1 | Model 2 | Model 3 | |||
|---|---|---|---|---|---|---|
| HR (95% CI) | p value | HR (95% CI) | p value | HR (95% CI) | p value | |
| Non-obesity | 1 (reference) | 1 (reference) | 1 (reference) | |||
| Obesity | 1.35(1.26,1.46) | < 0.001 | 1.34(1.25,1.45) | < 0.001 | 1.36(1.26,1.47) | < 0.001 |
Model 1: no adjusted
Model 2: adjusted for demographic variables (age, sex, marital status, education attainment)
Model 3: Model 2 + smoking and drinking status
Abbreviations: CI confidence intervals, CVD cardiovascular disease, HR hazard ratio
Association between obesity trajectories and CVD
Table 4 showed the relationship between obesity trajectories and CVD risk. After adjusting for age, sex, marital status, education attainment, smoking status, and drinking status, individuals in the consistent obesity group had a greater likelihood of CVD (HR: 1.50; 95% CI: 1.30–1.72; p < 0.001) compared to the no obesity group (reference group). Similarly, those with increasing obesity also showed an increased risk (HR: 1.34; 95% CI: 1.11–1.62; p = 0.003). Conversely, the groups with decreasing obesity (HR: 0.97; 95% CI: 0.54–1.75; p = 0.320) and fluctuating obesity (HR; 1.18, 95% CI: 0.90–1.55; p = 0.239) were not significantly correlated with a higher risk of CVD. These p values tested the null hypothesis that the HR equals 1, with values below 0.050 indicating statistically significant associations.
Table 4.
Association between obesity trajectories and risk of CVD
| Obesity trajectories | Model 1 | Model 2 | Model 3 | |||
|---|---|---|---|---|---|---|
| HR (95% CI) | p value | HR (95% CI) | p value | HR (95% CI) | p value | |
| No obesity | 1 (reference) | 1 (reference) | 1 (reference) | |||
| Decreasing obesity | 0.93(0.51,1.70) | 0.312 | 0.97(0.53,1.75) | 0.323 | 0.97(0.54,1.75) | 0.320 |
| Fluctuating obesity | 1.18(0.90,1.56) | 0.225 | 1.18(0.90,1.55) | 0.244 | 1.18(0.90,1.55) | 0.239 |
| Increasing obesity | 1.29(1.07,1.56) | 0.008 | 1.33(1.10,1.60) | 0.003 | 1.34(1.11,1.62) | 0.003 |
| Consistent obesity | 1.44(1.26,1.66) | < 0.001 | 1.46(1.27,1.68) | < 0.001 | 1.50(1.30,1.72) | < 0.001 |
Model 1: no adjusted
Model 2: adjusted for demographic variables (age, sex, marital status, education attainment)
Model 3: Model 2 + smoking and drinking status
Abbreviations: CI confidence intervals, CVD cardiovascular disease, HR hazard ratio
Subgroup and sensitivity analyses
Consistent findings were observed in analyses with stratification by age, sex, marital status, education attainment, smoking status, and drinking status (Table 5). All interaction p values were > 0.050, indicating no evidence of interaction. Additionally, the sensitivity analysis indicated similar trends when heart disease and stroke were analyzed as separate outcomes (Table S6). When analyses were stratified by cohort, consistent patterns were observed in both CHARLS and ELSA (Table S7). Among the 7,154 participants included in the trajectory analysis, 4,524 were from CHARLS and 2,630 from ELSA. In CHARLS, consistent obesity (HR: 1.89; 95% CI: 1.59–2.24; p < 0.001) and increasing obesity (HR: 1.69; 95% CI: 1.35–2.12; p < 0.001) were significantly associated with CVD risk. In ELSA, similar associations were found for consistent obesity (HR: 1.34; 95% CI: 1.21–1.52; p < 0.001) and increasing obesity (HR: 1.22; 95% CI: 1.14–1.40; p = 0.002). Decreasing and fluctuating trajectories were not significantly associated with CVD in either cohort. Besides, in sensitivity analyses restricted to participants with BMI ≥ 30 kg/m², adjustment for comorbidities had minimal impact on the association with CVD risk (Table S8). This minimal change suggests that any residual confounding by these conditions is negligible and does not inflate our primary effect estimates. For the competing risk analysis, after multiple adjustments for covariates, the comparative analysis revealed that the estimates for the obesity trajectories and CVD association were closely aligned between the competing risk and Cox proportional hazards models, with the maximum difference between the HR and the subdistribution hazard ratio was 0.07, and the maximum HR/SHR ratio was 1.08 (Table S9).
Table 5.
Subgroup analyses for the association between obesity trajectories and CVD
| Characteristic | No obesity | Decreasing obesity | Fluctuating obesity | Increasing obesity | Consistent obesity | p interaction |
|---|---|---|---|---|---|---|
| Age | 0.089 | |||||
| < 60 | 1 (reference) | 0.94(0.52,1.49) | 1.08(0.70,1.67) | 1.65(1.26,2.15) | 1.96(1.60,2.41) | |
| ≥ 60 | 1 (reference) | 1.06(0.69,1.58) | 1.24(0.87,1.77) | 1.09(0.83,1.43) | 1.14(1.03,1.39) | |
| Sex | 0.164 | |||||
| Female | 1 (reference) | 0.94(0.59,1.41) | 1.12(0.78,1.63) | 1.29(0.99,1.68) | 1.49(1.23,1.81) | |
| Male | 1 (reference) | 1.09(0.72,1.66) | 1.26(0.84,1.90) | 1.39(1.05,1.83) | 1.49(1.20,1.84) | |
| Marital status | 0.427 | |||||
| Married or partnered | 1 (reference) | 1.19(0.82,1.46) | 1.26(0.94,1.69) | 1.49(1.21,1.84) | 1.67(1.43,1.95) | |
| Other marital status | 1 (reference) | 0.84(0.40,1.37) | 0.88(0.42,1.84) | 0.80(0.51,1.27) | 1.21(1.06,1.42) | |
| Education attainment | 0.186 | |||||
| Below high school | 1 (reference) | 0.93(0.51,1.70) | 1.15(0.84,1.56) | 1.31(1.04,1.63) | 1.48(1.26,1.74) | |
| High school | 1 (reference) | 1.20(0.72,2.15) | 1.37(0.61,3.06) | 1.41(0.89,2.24) | 1.43(1.02,2.09) | |
| College or above | 1 (reference) | 1.18(0.67,2.03) | 1.36(0.59,3.13) | 1.61(0.87,2.96) | 2.04(1.31,3.19) | |
| Smoking status | 0.271 | |||||
| Non-current smoker | 1 (reference) | 0.98(0.59,1.82) | 1.20(0.87,1.64) | 1.42(1.14,1.77) | 1.50(1.27,1.77) | |
| Current smoker | 1 (reference) | 0.92(0.46,1.69) | 1.18(0.68,2.05) | 1.14(0.77,1.70) | 1.59(1.18,2.13) | |
| Alcohol status | 0.364 | |||||
| Non-current drinker | 1 (reference) | 0.90(0.41,1.98) | 1.11(0.75,1.62) | 1.37(1.04,1.80) | 1.65(1.36,2.02) | |
| Current drinker | 1 (reference) | 1.04(0.59,2.11) | 1.28(0.86,1.89) | 1.31(1.01,1.71) | 1.38(1.13,1.69) |
Values are hazard ratios (95% confidence interval). Each stratification was adjusted for age, sex, marital status, education attainment, smoking status, and drinking status except the stratification factor itself
Abbreviations: CVD cardiovascular disease
Added value of the EASO framework over BMI-based definition
Compared to the traditional BMI-based definition (16.40% obesity prevalence), the EASO framework classified 43.32% of participants as having obesity, reclassifying 26.92% (n = 3,192) from the overweight BMI category (Table S10). In terms of predictive performance, the EASO framework showed stronger association with CVD risk (HR: 1.36, 95% CI: 1.26–1.47; p < 0.001) than BMI-based obesity (HR: 1.24, 95% CI: 1.14–1.36; p < 0.001). Discrimination improved modestly (area under the curve: 0.60 vs. 0.51), and the continuous NRI was 0.18 (95% CI: 0.17–0.20), indicating significant improvement in risk classification (Table S11).
Discussion
In this study involving two prospective cohorts, we explored the associations between the new framework definition of obesity and CVD risk in middle-aged and older individuals from both static and dynamic viewpoints. Results from the static single-node follow-up study showed that compared with participants without baseline obesity, those with baseline obesity had a 36% higher risk of developing CVD. In the dynamic analyses, five obesity trajectories were determined, including no obesity, decreasing obesity, fluctuating obesity, increasing obesity, and consistent obesity. Compared to their counterparts in the no obesity group, participants in the consistent obesity and increasing obesity group had 50% and 34% higher prevalence of CVD, respectively. In contrast, the decreasing obesity and fluctuating obesity were not significantly associated with CVD risk.
Obesity is a serious healthcare problem worldwide. According to the World Health Organization, in 2022, 16% of adults aged 18 years and over were obesity, and the global prevalence of obesity will reach 17% in men and 22% in women by 2030 [23]. With its multiple complications, obesity has led to rising concerns about the worsening situation. Research indicated that obesity has a significant impact on both physical and mental health, contributing to CVD, diabetes, respiratory disease, cancers, as well as sleep disruptions, mobility problems, and mental distress [24, 25]. Currently, the diagnosis of obesity in both clinical practice and epidemiological research predominantly relies on BMI [7, 26]. The WHO defines obesity as a BMI greater than or equal to 30 kg/m2. However, the reliance on BMI alone as the diagnostic criterion is increasingly recognized as inadequate for comprehensive risk assessment. A primary limitation of BMI is its inability to differentiate between adipose tissue and lean mass or to account for critical differences in body fat distribution, particularly visceral adiposity [27]. Consequently, BMI can lead to significant clinical misclassification. It may underestimate adiposity and related health risks in individuals with a BMI in the normal or overweight range who possess excess body fat—a scenario common in older adults, those with sarcopenia, or certain ethnic groups [28]. Conversely, BMI can overdiagnose obesity in muscular individuals, such as athletes, whose elevated BMI stems from lean mass rather than excess adiposity [29].
Anthropometric measures of central adiposity, such as the WHtR, are superior to BMI alone in predicting visceral fat and cardiovascular risk [30]. Although their incorporation can improve the identification of excess adiposity and cardiometabolic risk, they still fall short of defining obesity as a disease. Anthropometric indices primarily reflect anatomical fat distribution but do not capture the presence or severity of obesity-driven organ dysfunction or clinical complications. More fundamentally, these measures, akin to BMI, provide no direct information on the functional impairments or specific medical comorbidities that constitute the core morbidity of the disease. Therefore, diagnostic approaches based solely on anthropometry, whether using BMI alone or in combination with other measures, risk both underdiagnosis and overdiagnosis of obesity-related illness [31]. This critical gap underscores the necessity for a more holistic diagnostic framework that moves beyond static anatomical assessments, necessitating an effective definition that integrates evidence of adiposity-related dysfunction. Accordingly, the EASO framework [8], adopted in this study, responds to this need by combining anthropometric criteria (BMI and WHtR) with the explicit assessment of medical, functional, or psychological complications to diagnose obesity as a clinical entity. This framework conceptualizes obesity as a chronic disease in which excess or dysfunctional adiposity manifests not only through anthropometric changes but also through its systemic health consequences. Each of these selected complications has well-established bidirectional relationships with obesity. Excess adiposity contributes to their pathogenesis through mechanisms such as chronic low-grade inflammation, insulin resistance, adipokine dysregulation, and mechanical load, while their presence, in turn, indicates that obesity has progressed from a risk factor to an active disease state with end-organ effects [8]. The clinical significance and potential impact of this integrated approach were underscored by a recent large-scale validation study [31]. Using data from 44,030 U.S. adults in the National Health and Nutrition Examination Survey 1999–2018, it was demonstrated that while the traditional BMI-based definition classified 35.4% of participants as having obesity, application of the EASO framework increased this proportion to 54.2%. This substantial reclassification of nearly one in five adults from the overweight to the obesity category highlights the framework’s enhanced sensitivity in identifying individuals with adiposity-related health risks, reinforcing the imperative to move beyond BMI-centric definitions in both research and clinical practice.
Extensive prior studies have established a robust association between obesity, defined at a single time point, and an elevated risk of CVD [32–36]. However, the dynamic nature of obesity, including its onset, progression, remission, and fluctuation over time, and how these longitudinal trajectories distinctly influence CVD risk remain largely unexplored. Our study addressed this gap by applying the EASO diagnostic framework across multiple waves of the CHARLS and ELSA cohorts to characterize five distinct obesity trajectories. We found that participants exhibiting a consistent or increasing obesity trajectory had a significantly higher prevalence of CVD (50% and 34% higher, respectively) compared to those who remained non-obese. In contrast, the decreasing and fluctuating obesity trajectory were not associated with a significantly elevated CVD risk. These findings underscored that the direction and persistence of adiposity status are critical determinants of cardiovascular health, moving beyond a static, cross-sectional assessment. Specifically, consistent and increasing obesity conferred substantially elevated risk, underscoring the clinical importance of early intervention to prevent obesity onset and progression. While our study did not demonstrate a significant protective effect of decreasing obesity, possibly due to metabolic legacy effects, insufficient weight loss magnitude, or limited statistical power, it reinforced the need for long-term weight management and regular monitoring of obesity status over time. Future research with larger samples, more frequent weight assessments, and longer follow-up is warranted to determine whether and under what conditions weight reduction translates into measurable cardiovascular benefit. Besides, the consistency of our findings across both CHARLS and ELSA cohorts strengthened the generalizability of our conclusions. Despite differences in population characteristics and obesity prevalence between China and the United Kingdom, the patterns of association between EASO definition of obesity trajectories and CVD risk remained remarkably similar across the combined 7,154 participants. This cross-cohort consistency suggests that the prognostic value of dynamic obesity assessment may be broadly applicable across diverse settings.
The differential risk associated with various obesity trajectories can be interpreted through several interconnected mechanisms. The heightened risk observed in the consistent obesity group likely reflects cumulative exposure to adverse metabolic and inflammatory milieus. Prolonged excess adiposity, particularly visceral fat, drives chronic low-grade inflammation, endothelial dysfunction, insulin resistance, and unfavorable lipid profiles [6, 37, 38], leading to sustained arterial damage over time [39]. The increasing obesity trajectory suggests a progressive pathological process, where escalating adiposity may accelerate these same mechanisms, surpassing a threshold of compensatory adaptation and triggering overt cardiovascular pathology [40]. In contrast, the non-significant association between a decreasing obesity trajectory and CVD risk in our study suggests that a reduction in adiposity status may mitigate the elevated cardiovascular risk typically associated with obesity. This observation aligns with the established concept that sustained weight loss can improve cardiometabolic risk factors, such as reducing blood pressure, improving lipid profiles, and enhancing insulin sensitivity, thereby potentially decelerating the progression of atherosclerosis [5, 41]. The lack of excess risk in the fluctuating group further suggests that variable weight patterns, while not optimal, may not confer the same cumulative biological insult as persistently high or escalating adiposity [42]. Future research measuring serial biomarkers (e.g., inflammatory cytokines, adipokines, imaging of vascular health) alongside trajectory groups is warranted to elucidate these pathways.
In our subgroup analyses, the associations between obesity trajectories and CVD risk were generally consistent across strata of age, sex, marital status, education attainment, smoking status, and drinking status, with no statistically significant interactions detected (all p for interaction > 0.050). Nevertheless, each of these factors may influence obesity-related cardiovascular risk through distinct pathways. Sex differences in adipose tissue distribution, with women typically having greater subcutaneous fat and men having greater visceral adiposity, may modulate metabolic consequences of obesity [43]. Education attainment, as a marker of socioeconomic position, may affect obesity prevalence and CVD outcomes through differential access to preventive healthcare, health literacy, and lifestyle behaviors [44]. Besides, smoking and alcohol consumption are established cardiovascular risk factors that may synergistically interact with obesity-related inflammation, oxidative stress, and endothelial dysfunction [45, 46]. The absence of significant interaction in our study may reflect the predominant biological impact of obesity itself, which may outweigh the modifying influence of these covariates. Future studies with larger, more diverse populations and repeated measurements of behavioral factors over time are warranted to further explore potential interactions and mediating pathways.
The clinical implications of our findings, derived from applying the novel EASO framework, underscored the necessity of integrating dynamic obesity trajectory assessment into CVD prevention strategies. The EASO criteria, which incorporate central adiposity and clinical comorbidities, enable a more precise identification of high-risk individuals beyond BMI alone. Therefore, clinicians should adopt this holistic assessment by regularly monitoring not only weight but also WHtR and the status of key comorbidities. For individuals identified as following a consistent or increasing obesity trajectory under this framework, who face a significantly elevated CVD risk, timely and intensive management is warranted. This may include structured lifestyle intervention, pharmacological therapy for weight management, and aggressive control of associated comorbidities. Conversely, for those showing a decreasing trajectory, clinical efforts should focus on reinforcing and supporting weight maintenance to preserve cardiovascular benefits.
Several limitations of this study should be acknowledged. First, the definition of CVD, based on self-reported physician diagnoses, lacks validation against clinical records and may underestimate asymptomatic or undiagnosed cases. Second, while the EASO framework represents a significant advance, the fixed WHtR cutoff and the selected list of complications may not be universally optimal across all populations or capture every relevant adiposity-related dysfunction. Third, despite adjusting for key covariates, residual confounding from unmeasured lifestyle, dietary, or genetic factors is possible. Furthermore, although our subgroup analyses suggested no significant interaction by age, sex, marital status, education attainment, smoking status, and drinking status, these analyses may have been underpowered to detect modest interaction effects. More detailed longitudinal assessment of these factors could provide deeper insights into their potential mediating or modifying roles in the obesity-CVD relationship. Fourth, although we observed no significant association between decreasing obesity and CVD risk, this finding should be interpreted with caution. The decreasing obesity group was the smallest in our trajectory analysis, limiting statistical power to detect modest protective effects. Moreover, our trajectory definition captured the presence or absence of obesity at three time points but did not quantify the magnitude and speed of weight loss, which are factors that may critically influence cardiovascular outcomes. Future studies with more granular weight data and larger sample sizes are needed to determine whether specific patterns of weight reduction confer cardiovascular protection. Fifth, despite efforts to minimize bias, excluding participants lost to follow-up may introduce selection bias. Our comparison of included and excluded participants revealed that those lost to follow-up had a higher-risk profile (older, male, unmarried, lower education), suggesting that our estimates may be conservative. Finally, as with any observational study, the associations identified do not confirm causality, although the longitudinal design strengthens temporal inference. These limitations highlight opportunities for future research to employ more intensive longitudinal data collection, medically confirmed clinical outcomes, and investigate the generalizability of the novel EASO framework across diverse ethnic and clinical populations.
Conclusion
In conclusion, this longitudinal study applied the EASO new framework to define obesity and identified distinct obesity trajectories across two nationally cohorts. We found that consistent and increasing obesity trajectories were significantly associated with a higher prevalence of CVD, whereas decreasing and fluctuating trajectories were not. These findings highlight that the dynamic course of adiposity, assessed through a holistic diagnostic definition, is a critical determinant of cardiovascular risk beyond a single measurement. They reinforce the clinical importance of early and sustained weight management and support the integration of trajectory-based risk assessment into preventive cardiology strategies. Future research should further validate the EASO criteria and investigate the mechanisms linking specific obesity pathways to cardiovascular outcomes.
Supplementary Information
Acknowledgements
We thank Home for Researchers (https://www.home-for-researchers.com/) for their linguistic assistance.
Authors’ contributions
Dingyuan Tu: Methodology, Formal analysis, Writing – original draft; Jinru Li: Formal analysis, Data curation; Jianming Wang: Methodology, Software, Conceptualization, Writing – review & editing.
Funding
This work was supported by Liaoning Province Union Program - Application Fundamental Research Project (2023JH2/101700138) and Shenyang Joint Logistics Support Force Young Elite Cultivation Program (2025SL0012).
Data availability
The original data for this study are available on their respective websites: the China Health and Retirement Longitudinal Study (http://charls.pku.edu.cn/index/en.html) and the English Longitudinal Study of Ageing (https://www.elsa-project.ac.uk).
Declarations
Ethics approval and consent to participate
The CHARLS was approved by the Biomedical Ethics Review Committee of Peking University (IRB00001052-11015). The ELSA was approved by the National Research Ethics Service (London Multicentre Research Ethics Committee, MREC/01/2/91, with subsequent approvals for each wave). All participants provided written informed consent at enrolment. The present study involved secondary analysis of deidentified data from these cohorts. No new data were collected, and no additional ethics approval was required. In compliance with the Declaration of Helsinki, this research was conducted.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Dingyuan Tu and Jinru Li contributed equally to this work.
References
- 1.Vaduganathan M, Mensah GA, Turco JV, Fuster V, Roth GA. The Global Burden of Cardiovascular Diseases and Risk. J Am Coll Cardiol. 2022;80(25):2361–71. [DOI] [PubMed] [Google Scholar]
- 2.Chong B, Jayabaskaran J, Jauhari SM, Chan SP, Goh R, Kueh MTW, Li H, Chin YH, Kong G, Anand VV et al. Global burden of cardiovascular diseases: projections from 2025 to 2050. Eur J Prev Cardiol. 2024 ;32(11):1001–15. [DOI] [PubMed]
- 3.Figtree GA, Vernon ST, Harmer JA, Gray MP, Arnott C, Bachour E, Barsha G, Brieger D, Brown A, Celermajer DS, et al. Clinical Pathway for Coronary Atherosclerosis in Patients Without Conventional Modifiable Risk Factors: JACC State-of-the-Art Review. J Am Coll Cardiol. 2023;82(13):1343–59. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Consortium GCR. Global Effect of Modifiable Risk Factors on Cardiovascular Disease and Mortality. N Engl J Med. 2023;389(14):1273–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Koskinas KC, Van Craenenbroeck EM, Antoniades C, Bluher M, Gorter TM, Hanssen H, Marx N, McDonagh TA, Mingrone G, Rosengren A, et al. Obesity and cardiovascular disease: an ESC clinical consensus statement. Eur Heart J. 2024;45(38):4063–98. [DOI] [PubMed] [Google Scholar]
- 6.Koenen M, Hill MA, Cohen P, Sowers JR. Obesity, Adipose Tissue and Vascular Dysfunction. Circ Res. 2021;128(7):951–68. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Sweatt K, Garvey WT, Martins C. Strengths and Limitations of BMI in the Diagnosis of Obesity: What is the Path Forward? Curr Obes Rep. 2024;13(3):584–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Busetto L, Dicker D, Fruhbeck G, Halford JCG, Sbraccia P, Yumuk V, Goossens GH. A new framework for the diagnosis, staging and management of obesity in adults. Nat Med. 2024;30(9):2395–9. [DOI] [PubMed] [Google Scholar]
- 9.Sbraccia P, Dicker D. Obesity is a chronic progressive relapsing disease of particular interest for internal medicine. Intern Emerg Med. 2023;18(1):1–5. [DOI] [PubMed] [Google Scholar]
- 10.Zhao Y, Hu Y, Smith JP, Strauss J, Yang G. Cohort profile: the China Health and Retirement Longitudinal Study (CHARLS). Int J Epidemiol. 2014;43(1):61–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Steptoe A, Breeze E, Banks J, Nazroo J. Cohort profile: the English longitudinal study of ageing. Int J Epidemiol. 2013;42(6):1640–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Palaniappan LP, Wong EC, Shin JJ, Fortmann SP, Lauderdale DS. Asian Americans have greater prevalence of metabolic syndrome despite lower body mass index. Int J Obes (Lond). 2011;35(3):393–400. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Jones DW, Ferdinand KC, Taler SJ, Johnson HM, Shimbo D, Abdalla M, Altieri MM, Bansal N, Bello NA, Bress AP et al. 2025 AHA/ACC/AANP/AAPA/ABC/ACCP/ACPM/AGS/AMA/ASPC/NMA/PCNA/SGIM Guideline for the prevention, detection, evaluation and management of high blood pressure in adults: a report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Hypertension. 2025;82(12):e350. [DOI] [PubMed]
- 14.ElSayed NA, McCoy RG, Aleppo G, Balapattabi K, Beverly EA, Briggs Early K, Bruemmer D, Ebekozien O, Echouffo-Tcheugui JB, Ekhlaspour L, et al. : 2. Diagnosis and Classification of Diabetes: Standards of Care in Diabetes—2025. Diabetes Care. 2025;48(Supplement1):S27–49. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Wu S, Yang X, Chen Y, Wang Y, Liang H, Xu W, Wang J, Shao X, Zhang H, Zhong Z et al. Changes in depressive symptoms as predictors of incident cardiovascular disease: insights from four prospective cohorts. Eur J Prev Cardiol. 2025. [DOI] [PubMed]
- 16.Y R, L X, X R, D X, L L, H X. F. C, J. Z: Association between internet exclusion and depressive symptoms among older adults: panel data analysis of five longitudinal cohort studies. EClinicalMedicine. 2024;75:102767. [DOI] [PMC free article] [PubMed]
- 17.He X, Li C, Hua G, Wang Y, Zhang L, Cheng C, Jiao J, Yu H, Yang X, Liu L. Cross-national differences in the association between estimated cardiorespiratory fitness and depressive symptoms among older adults: findings from three nationwide cohort studies. BMC Med. 2025;23(1):363. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Yao X, Leng Y, Cao J, Yu M, Zhou Y. Eight-year chronic pain trajectory and risk of cardiovascular disease: Evidence from 10 years of multicenter follow-up. J Pain. 2025;34:105478. [DOI] [PubMed] [Google Scholar]
- 19.Zhang D, Shen C, Chen N, Liu C, Hu J, Lau KK, Wen Z, Qiu A. Long-term obesity impacts brain morphology, functional connectivity and cognition in adults. Nat Mental Health. 2025;3(4):466–78. [Google Scholar]
- 20.Cheng F, Wang Z, Li Y, Wang P, Wang C. Longitudinal associations of chronic pain severity trajectories and number of pain site trajectories with risk of limitations in ability in daily activities: evidence from two 10-year prospective cohort studies. J Nutr Health Aging. 2026;30(3):100781. [DOI] [PMC free article] [PubMed]
- 21.Zhang Z, Xu H, Zhang R, Yan Y, Ling X, Meng Y, Zhang X, Wang Y. Frailty and depressive symptoms in relation to cardiovascular disease risk in middle-aged and older adults. Nat Commun. 2025;16(1):6008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Li J, Wei X. Baseline and changes in cardiometabolic index and incident cardiovascular disease in two prospective cohorts. Am J Prev Cardiol. 2025;23:101046. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Obesity and overweight. https://www.who.int/news-room/fact-sheets/detail/obesity-and-overweight.
- 24.Xiao N, Ding Y, Cui B, Li R-r, Qu X-r, Zhou H, Au K-h, Fan X-d, Xie J-c, Huang Y et al. Navigating obesity: a comprehensive review of epidemiology, pathophysiology, complications and management strategies. Innov Med. 2024;2(3).
- 25.Banerjee D, Mani A. Obesity’s systemic impact: exploring molecular and physiological links to diabetes, cardiovascular disease, and heart failure. Front Endocrinol (Lausanne). 2025;16:1681766. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Bluher M. Obesity: global epidemiology and pathogenesis. Nat Rev Endocrinol. 2019;15(5):288–98. [DOI] [PubMed] [Google Scholar]
- 27.Yang L, Ge Y, Zhu Q, Zhang Q, Wang L, Wang X, Yang Y, Shi H. The Combination of Fat Distribution and BMI Redefines Obesity: Result From NHANES. J Cachexia Sarcopenia Muscle. 2025;16(4):e70013. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Bray GA. Beyond BMI. Nutrients. 2023;15(10):2254. [DOI] [PMC free article] [PubMed]
- 29.Lee DH, Keum N, Hu FB, Orav EJ, Rimm EB, Willett WC, Giovannucci EL. Predicted lean body mass, fat mass, and all cause and cause specific mortality in men: prospective US cohort study. BMJ. 2018;362:k2575. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Abdi Dezfouli R, Mohammadian Khonsari N, Hosseinpour A, Asadi S, Ejtahed HS, Qorbani M. Waist to height ratio as a simple tool for predicting mortality: a systematic review and meta-analysis. Int J Obes (Lond). 2023;47(12):1286–301. [DOI] [PubMed] [Google Scholar]
- 31.Dicker D, Karpati T, Promislow S, Reges O. Implications of the European Association for the Study of Obesity’s New Framework Definition of Obesity: Prevalence and Association With All-Cause Mortality. Ann Intern Med. 2025;178(8):1065–72. [DOI] [PubMed] [Google Scholar]
- 32.Jia X, Hu C, Xu Y, Yin Y, Lin H, Zheng R, Li M, Xu M, Wang T, Zhao Z, et al. Revisiting obesity thresholds for cardiovascular disease and mortality risk in Chinese adults: Age- and gender-specific insights. Cell Rep Med. 2025;6(9):102309. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Lyall DM, Celis-Morales C, Ward J, Iliodromiti S, Anderson JJ, Gill JMR, Smith DJ, Ntuk UE, Mackay DF, Holmes MV, et al. Association of Body Mass Index With Cardiometabolic Disease in the UK Biobank: A Mendelian Randomization Study. JAMA Cardiol. 2017;2(8):882–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Holmes MV, Lange LA, Palmer T, Lanktree MB, North KE, Almoguera B, Buxbaum S, Chandrupatla HR, Elbers CC, Guo Y, et al. Causal effects of body mass index on cardiometabolic traits and events: a Mendelian randomization analysis. Am J Hum Genet. 2014;94(2):198–208. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Nordestgaard BG, Palmer TM, Benn M, Zacho J, Tybjaerg-Hansen A, Davey Smith G, Timpson NJ. The effect of elevated body mass index on ischemic heart disease risk: causal estimates from a Mendelian randomisation approach. PLoS Med. 2012;9(5):e1001212. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Riaz H, Khan MS, Siddiqi TJ, Usman MS, Shah N, Goyal A, Khan SS, Mookadam F, Krasuski RA, Ahmed H. Association Between Obesity and Cardiovascular Outcomes: A Systematic Review and Meta-analysis of Mendelian Randomization Studies. JAMA Netw Open. 2018;1(7):e183788. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Kolb H. Obese visceral fat tissue inflammation: from protective to detrimental? BMC Med. 2022;20(1):494. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Saponaro C, Sabatini S, Gaggini M, Carli F, Rosso C, Positano V, Armandi A, Caviglia GP, Faletti R, Bugianesi E, et al. Adipose tissue dysfunction and visceral fat are associated with hepatic insulin resistance and severity of NASH even in lean individuals. Liver Int. 2022;42(11):2418–27. [DOI] [PubMed] [Google Scholar]
- 39.Powell-Wiley TM, Poirier P, Burke LE, Despres JP, Gordon-Larsen P, Lavie CJ, Lear SA, Ndumele CE, Neeland IJ, Sanders P, et al. Obesity and Cardiovascular Disease: A Scientific Statement From the American Heart Association. Circulation. 2021;143(21):e984–1010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Kong Y, Yang H, Nie R, Zhang X, Zuo F, Zhang H, Nian X. Obesity: pathophysiology and therapeutic interventions. Mol Biomed. 2025;6(1):25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Ma C, Avenell A, Bolland M, Hudson J, Stewart F, Robertson C, Sharma P, Fraser C, MacLennan G. Effects of weight loss interventions for adults who are obese on mortality, cardiovascular disease, and cancer: systematic review and meta-analysis. BMJ. 2017;359:j4849. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Oh TJ, Moon JH, Choi SH, Lim S, Park KS, Cho NH, Jang HC. Body-Weight Fluctuation and Incident Diabetes Mellitus, Cardiovascular Disease, and Mortality: A 16-Year Prospective Cohort Study. J Clin Endocrinol Metab. 2019;104(3):639–46. [DOI] [PubMed] [Google Scholar]
- 43.Torres SV, Man K, Elmzzahi T, Malko D, Chisanga D, Liao Y, Prout M, Abbott CA, Tang A, Wu J, et al. Two regulatory T cell populations in the visceral adipose tissue shape systemic metabolism. Nat Immunol. 2024;25(3):496–511. [DOI] [PubMed] [Google Scholar]
- 44.Z LWB, Z Z, Y L, Z M, J Y, L Y, Z M, W L. Body-mass index and obesity in urban and rural China: findings from consecutive nationally representative surveys during 2004-18. Lancet. 2021;398(10294):53–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.China Kadoorie Biobank Collaborative G. Joint impact of polygenic risk score and lifestyles on early- and late-onset cardiovascular diseases. Nat Hum Behav. 2024;8(9):1810–8. [DOI] [PubMed] [Google Scholar]
- 46.K DLCLJAA, C AMSMEHSEJSA, V RS, B EJ et al. PM. C : Prospective associations of obesity and obesity severity with 9 cardiovascular outcomes: the cross-cohort collaboration. Circulation. 2026. [DOI] [PMC free article] [PubMed]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The original data for this study are available on their respective websites: the China Health and Retirement Longitudinal Study (http://charls.pku.edu.cn/index/en.html) and the English Longitudinal Study of Ageing (https://www.elsa-project.ac.uk).