Abstract
Background
Cardiovascular–kidney–metabolic (CKM) syndrome underscores the interlinked pathophysiology of metabolic dysregulation, chronic kidney disease, and cardiovascular disease (CVD). Although depression is a key psychosocial factor, evidence linking its longitudinal trajectories to incident CVD among middle-aged and older adults with CKM stages 0–3 remains limited.
Methods
We conducted a prospective cohort study using data from the China Health and Retirement Longitudinal Study (CHARLS) collected between 2011 and 2020, including participants at CKM stages 0–3. Depressive symptoms were assessed using the Center for Epidemiologic Studies Depression Scale (CES-D) across the 2011, 2013, and 2015 waves. Group-based trajectory modeling (GBTM) was applied to identify depressive symptom trajectories. Multivariable Cox proportional hazards models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for incident CVD (including heart disease and stroke).
Results
Among 6,152 participants (49% male, median age 58.0 years), 1,080 incident CVD events occurred during the follow-up. We identified three distinct depressive symptom trajectories: Low-stable, Moderate-stable, and Persistently-high. The cumulative incidence of CVD across these groups was 13.0%, 17.3%, and 23.1%, respectively (P < 0.001). Relative to the Low-stable group, fully adjusted Cox models revealed a significantly higher risk of incident CVD in both the Moderate-stable (HR = 1.36; 95% CI: 1.16–1.60) and Persistently-high (HR = 1.88; 95% CI: 1.57–2.24) groups. Consistent risk elevations were observed for coronary heart disease and stroke analyzed separately.
Conclusion
In Chinese middle-aged and older adults with CKM stages 0–3, depressive symptom trajectories are heterogeneous. Both Persistently-high and Moderate-stable trajectories of elevated depressive symptoms are independently associated with an increased risk of incident CVD.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12872-026-06068-1.
Keywords: Cardiovascular–kidney–metabolic syndrome, Depressive symptoms, Group-based trajectory modeling, Cardiovascular disease, Middle-aged and older adults
Background
Cardiovascular–kidney–metabolic syndrome (CKM) is a clinical construct that integrates metabolic risk factors, chronic kidney disease (CKD), and cardiovascular disease (CVD) [1]. The American Heart Association (AHA) proposed this framework to emphasize the close, bidirectional links among metabolic dysregulation, kidney disease, and cardiovascular pathology, which ultimately increase the risk of multiorgan dysfunction and adverse cardiovascular outcomes [2, 3]. CKM syndrome is currently classified into stages 0 to 4 based on risk factors and diagnosed diseases. Stages 0 to 3 represent the preclinical phase, occurring before the onset of overt cardiovascular diseases, during which CVD prevention should be prioritized [3, 4].
Existing studies indicate that CKM risk arises from genetic susceptibility and traditional exposures such as overweight/obesity [5], unhealthy diet [6], and physical inactivity [7]. However, the contribution of modifiable psychosocial factors, particularly depression, remains underexplored in CKM. Depression is a major global health burden, affecting approximately 350 million people worldwide, and represents a highly common and significant health challenge among middle-aged and older adults [8, 9]. Recent research has shown that depressive symptoms are significantly associated with an increased all-cause mortality rate among participants with CKM stages 1 to 3, and this association is mediated by the systemic inflammatory response index [10]. He et al. [11] also found that higher levels of depressive severity are independently linked to the progression of CKM and an increased risk of death, a finding confirmed in data from two large nationally representative cohorts. These findings underscore the importance of incorporating mental health care into the management of CKM.
Most previous research has relied on cross-sectional designs or single-time-point assessments, overlooking the dynamic progression of depressive symptoms. In reality, depression follows heterogeneous trajectories, with some individuals experiencing temporary remission while others undergo prolonged or recurrent episodes. One study identified three distinct developmental trajectories of depressive symptoms among Chinese junior high school students, demonstrating that these pathways are shaped by dynamic interactions among individual characteristics, family environment, and behavioral factors [12]. Similarly, another study employed group-based trajectory modeling to characterize heterogeneous patterns of depressive symptoms in older adults, classifying their trajectories into three types: stable low, decline followed by an increase and continuously rising [13]. However, evidence remains limited on how these depressive trajectories impact the progression of CKM at various stages and influence cardiovascular outcomes.
Using repeated measurements from the China Health and Retirement Longitudinal Study (CHARLS), this study aimed to identify heterogeneous trajectories of depressive symptoms and to evaluate the associations of these trajectories with the risk of incident cardiovascular events—including heart disease, stroke, and their composite endpoint—among middle-aged and older individuals across CKM stages 0–3.
Materials and methods
Data source and ethics
We used de-identified data from the CHARLS, a nationally representative longitudinal survey administered by the National School of Development at Peking University. The target population for this study consisted of middle-aged and older adults (aged > = 45 years at baseline) enrolled in the CHARLS cohort. Details of the study design and data access are available on the official website (http://charls.pku.edu.cn/en) [14]. CHARLS was approved by the Peking University Biomedical Ethics Review Committee (IRB No. IRB00001052-11015). All participants provided written informed consent, including permission for secondary analyses.
Assessment of depressive symptoms
Depressive symptoms were assessed with the 10-item Center for Epidemiologic Studies Depression Scale (CES-D) [15]. Each item has a 4-level frequency response referring to the past week: rarely or none of the time (< 1 day), some or a little of the time (1–2 days), occasionally or a moderate amount of time (3–4 days), and most or all of the time (5–7 days). Total scores range from 0 to 30, with higher scores indicating more severe symptoms. The CES-D data used in our study were collected from the CHARLS surveys conducted during the 2011, 2013, and 2015 waves.
Definition and staging of CKM
According to the AHA presidential advisory [2], CKM syndrome is staged from 0 to 4 as follows: Stage 0: No CKM risk factors. Stage 1: Early metabolic disorders, such as overweight, abdominal obesity, and impaired glucose tolerance. Stage 2: Overt metabolic diseases, such as type 2 diabetes, hypertension, hypertriglyceridemia, and/or kidney disease. Stage 3: Subclinical cardiovascular involvement or individuals at high risk for cardiovascular disease, including those with elevated 10-year CVD risk and very high-risk CKD. CVD risk is assessed using the Framingham Risk Score, with a > 20% 10-year risk considered high. CKD staging follows KDIGO guidelines, with high-risk CKD defined as an eGFR < 30 ml/min/1.73 m². Stage 4: Established clinical CVD, including coronary heart disease, heart failure, stroke, peripheral artery disease, or atrial fibrillation. The detailed operational definitions adapting these AHA criteria to the specific variables available in the CHARLS cohort are comprehensively outlined in Table S1.
Assessment of incident CVD outcome
In the CHARLS, the diagnosis of CVD was based on self-reported doctor diagnoses by participants. The researchers asked participants if they had ever been diagnosed by a doctor with heart disease (including heart attacks, coronary heart disease, angina pectoris, or congestive heart failure) or stroke. For this study, newly reported cases of CVD in the 2018 and 2020 waves are considered the primary outcome. Given the periodic nature of the assessments, the exact date of CVD onset was inherently interval‑censored. To address this, the event time for incident cases was imputed as the midpoint between the date of the last documented CVD‑free interview and the date of the interview at which the initial diagnosis was reported. Participants who did not develop CVD were right‑censored at the date of their final available interview. To ensure a rigorous prospective sequence and minimize reverse causality, the 2011–2015 period served as the CES-D measurement window, and participants were strictly required to be free of CVD throughout this entire period.
Data collection
Baseline information was collected by trained interviewers using standardized questionnaires. Categorical variables included sex, education level, residence, smoking status, drinking status, hypertension, diabetes, dyslipidemia, CKM stage, and use of lipid-lowering, glucose-lowering, and antihypertensive medications. Continuous variables included age, height, weight, waist circumference, systolic blood pressure (SBP), diastolic blood pressure (DBP), total cholesterol (TC), high-density lipoprotein (HDL), low-density lipoprotein (LDL), triglycerides (TG), fasting glucose, serum creatinine, and uric acid.
Statistical analysis
Using CES-D-10 scores from 2011, 2013, and 2015, we applied group-based trajectory modeling (GBTM) to identify depressive-symptom trajectories. The trajectory analysis was performed using the gbmt package in R. To ensure methodological transparency and reproducibility, the core analytical code is provided in Supplementary material 1. Model fit was evaluated using multiple information criteria—Akaike information criterion (AIC), adjusted AIC (aAIC), Bayesian information criterion (BIC), consistent AIC (CAIC), sample-size–adjusted BIC (SS-BIC), and the Hannan–Quinn information criterion (HQIC)—with lower values indicating better fit and classification. To ensure robust classification, the average posterior probability (AvePP) of assignment was required to be > 0.70 for all groups. Furthermore, to maintain adequate statistical power and representativeness, each trajectory class was required to comprise at least 5% of the total analytical sample. Ultimately, the final trajectory categories were determined by taking both rigorous statistical criteria and clinical meaningfulness into comprehensive consideration [16, 17].
Baseline characteristics are presented as medians (interquartile ranges [IQRs]) for continuous variables with skewed distributions and as counts (percentages) for categorical variables. The distribution of missing data for baseline covariates is detailed in Table S2. Missing values were imputed using a non-parametric random forest imputation algorithm via the missForest package. We fitted multivariable Cox proportional hazards models to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for the associations of depressive-symptom trajectories with incident CVD, coronary heart disease, and stroke. Covariate selection was guided by prior literature and clinical relevance [18, 19]. Models were sequentially adjusted for clinically relevant confounders: Model 1 was unadjusted; Model 2 adjusted for age, sex, education level, residence, waist circumference, smoking status, and drinking status; and Model 3 further adjusted for hypertension, diabetes, dyslipidemia, serum creatinine, and uric acid. The proportional hazards assumption was verified using Schoenfeld residuals, and multicollinearity was assessed using variance inflation factors (VIFs). Prespecified subgroup analyses stratified by sex, hypertension, diabetes, dyslipidemia, and CKM stage assessed consistency. In sensitivity analyses, a complete-case analysis was performed by excluding participants with missing data to test the robustness of our imputation approach. Additionally, to bypass potential midpoint imputation bias for interval-censored CVD events, we validated our primary results using a discrete-time survival model (with a complementary log-log link) treating survey waves as the discrete time unit. Two-sided P < 0.05 was considered statistically significant. All analyses were conducted in R version 4.4.0.
Results
Depressive symptom trajectories
The final analysis included 6,152 participants, based on the established inclusion and exclusion criteria (Fig. 1). Table S3 and Figure S1 present the fit indices for six candidate GBTM models of CES-D scores across the 2011, 2013, and 2015 surveys. By comprehensively evaluating the statistical fit indices, classification adequacy (AvePP > 0.70), and clinical practicality, we ultimately selected the three-trajectory model as the optimal fit. This model provided the most distinct and parsimonious categorization of longitudinal depressive burdens. As illustrated in Fig. 2, the identified trajectories were: Low-stable, characterized by persistently low CES-D scores (< 4); Moderate-stable, exhibiting moderately elevated scores stabilizing between 6 and 8; and Persistently-high, defined by consistently severe CES-D scores maintained at a distinctly high level throughout the follow-up period.
Fig. 1.
Flowchart of study participants
Fig. 2.
Longitudinal trajectories of depressive symptoms among participants from 2011 to 2015. Note: Three distinct trajectories of depressive symptoms were identified using GBTM across three consecutive CHARLS waves (2011, 2013, and 2015). The y-axis shows the CES-D score. Solid lines represent the mean predicted trajectories for the "Persistently-high", "Moderate-stable", and "Low-stable" subgroups
Characteristics of participants
Table 1 summarizes the baseline characteristics of the 6,152 participants (49% male), with a median age of 58.0 years (IQR, 51.0–64.0). Statistically significant differences (all P < 0.05) were observed across the depressive symptom trajectory groups for sex, age, education level, residence, smoking status, drinking status, dyslipidemia, height, weight, waist circumference, DBP, serum creatinine, uric acid, TC, and HDL-C. Compared to the Low-stable group, participants in the Persistently-high group were older, comprised a significantly higher proportion of females and rural residents, and had lower education levels. Furthermore, this high-risk trajectory group exhibited lower proportions of current smokers and alcohol consumers. In terms of physical and clinical profiles, they presented with noticeably lower median height, weight, waist circumference, and DBP. Interestingly, despite having lower baseline levels of serum creatinine and uric acid, this group demonstrated a striking elevation in specific lipid profiles, exhibiting significantly higher median concentrations of both TC and HDL-C.
Table 1.
Baseline characteristics of participants
| Characteristic | Overall N = 6,152 |
Trajectories of depressive symptoms | p-value | ||
|---|---|---|---|---|---|
| Low-stable N = 1,689 |
Moderate-stable N = 2,937 |
Persistently-high N = 1,526 |
|||
| Sex, n (%) | < 0.001 | ||||
| female | 3,144 (51) | 665 (39) | 1,494 (51) | 985 (65) | |
| male | 3,008 (49) | 1,024 (61) | 1,443 (49) | 541 (35) | |
| Age (years) | 58.00 (51.00, 64.00) | 57.00 (50.00, 63.00) | 58.00 (51.00, 64.00) | 58.00 (53.00, 65.00) | < 0.001 |
| Education, n (%) | < 0.001 | ||||
| higher than primary school | 1,908 (31) | 740 (44) | 874 (30) | 294 (19) | |
| primary school or lower | 4,244 (69) | 949 (56) | 2,063 (70) | 1,232 (81) | |
| Residence, n (%) | < 0.001 | ||||
| rural | 4,106 (67) | 947 (56) | 1,992 (68) | 1,167 (76) | |
| urban | 2,046 (33) | 742 (44) | 945 (32) | 359 (24) | |
| Smoking status, n (%) | < 0.001 | ||||
| no | 4,107 (67) | 1,035 (61) | 1,959 (67) | 1,113 (73) | |
| yes | 2,045 (33) | 654 (39) | 978 (33) | 413 (27) | |
| Drinking status , n (%) | < 0.001 | ||||
| no | 3,959 (64) | 971 (57) | 1,902 (65) | 1,086 (71) | |
| yes | 2,193 (36) | 718 (43) | 1,035 (35) | 440 (29) | |
| Hypertension, n (%) | 0.16 | ||||
| no | 4,877 (79) | 1,366 (81) | 2,311 (79) | 1,200 (79) | |
| yes | 1,275 (21) | 323 (19) | 626 (21) | 326 (21) | |
| Dyslipidemia, n (%) | 0.034 | ||||
| no | 5,725 (93) | 1,551 (92) | 2,756 (94) | 1,418 (93) | |
| yes | 427 (6.9) | 138 (8.2) | 181 (6.2) | 108 (7.1) | |
| Diabetes, n (%) | 0.073 | ||||
| no | 5,873 (95) | 1,624 (96) | 2,807 (96) | 1,442 (94) | |
| yes | 279 (4.5) | 65 (3.8) | 130 (4.4) | 84 (5.5) | |
| CKM_stage, n (%) | 0.25 | ||||
| Stage 0 | 426 (6.9) | 98 (5.8) | 211 (7.2) | 117 (7.7) | |
| Stage 1 | 1,144 (19) | 314 (19) | 551 (19) | 279 (18) | |
| Stage 2 | 2,359 (38) | 638 (38) | 1,118 (38) | 603 (40) | |
| Stage 3 | 2,223 (36) | 639 (38) | 1,057 (36) | 527 (35) | |
| Antihyperlipidemic agents, n (%) | 0.55 | ||||
| no | 5,930 (96) | 1,621 (96) | 2,835 (97) | 1,474 (97) | |
| yes | 222 (3.6) | 68 (4.0) | 102 (3.5) | 52 (3.4) | |
| Antidiabetic agents, n (%) | 0.64 | ||||
| no | 5,971 (97) | 1,643 (97) | 2,852 (97) | 1,476 (97) | |
| yes | 181 (2.9) | 46 (2.7) | 85 (2.9) | 50 (3.3) | |
| Antihypertensive agents, n (%) | 0.62 | ||||
| no | 5,222 (85) | 1,446 (86) | 2,485 (85) | 1,291 (85) | |
| yes | 930 (15) | 243 (14) | 452 (15) | 235 (15) | |
| Height (m) | 1.58 (1.52, 1.64) | 1.61 (1.55, 1.67) | 1.58 (1.53, 1.64) | 1.55 (1.50, 1.61) | < 0.001 |
| Weight (kg) | 58.00 (51.58, 65.50) | 60.90 (54.10, 68.30) | 57.90 (51.50, 65.20) | 55.30 (49.50, 62.20) | < 0.001 |
| Waist circumference(cm) | 84.40 (78.00, 91.10) | 85.50 (79.00, 92.40) | 84.38 (78.00, 91.00) | 83.50 (77.00, 90.20) | < 0.001 |
| SBP (mmHg) | 127.27 (115.33, 141.00) | 127.67 (116.33, 141.17) | 127.33 (115.67, 141.33) | 126.33 (113.67, 140.67) | 0.10 |
| DBP (mmHg) | 75.00 (68.00, 83.00) | 75.67 (68.33, 83.67) | 75.00 (68.27, 83.13) | 74.00 (66.67, 82.33) | < 0.001 |
| Glucose (mg/dL) | 102.78 (95.04, 113.22) | 103.50 (95.22, 113.04) | 102.60 (95.22, 113.58) | 102.42 (94.50, 112.68) | 0.31 |
| Creatinine (mg/dL) | 0.76 (0.66, 0.88) | 0.78 (0.68, 0.90) | 0.76 (0.66, 0.88) | 0.72 (0.63, 0.84) | < 0.001 |
| Uric acid (mg/dL) | 4.32 (3.59, 5.17) | 4.50 (3.73, 5.41) | 4.34 (3.62, 5.18) | 4.12 (3.44, 4.91) | < 0.001 |
| TC (mg/dL) | 190.98 (168.56, 214.95) | 188.66 (167.40, 213.02) | 191.75 (168.94, 215.34) | 191.37 (168.94, 217.27) | 0.018 |
| HDL (mg/dL) | 49.76 (40.59, 59.92) | 47.94 (39.43, 58.38) | 49.87 (40.98, 59.92) | 51.03 (41.37, 61.08) | < 0.001 |
| LDL (mg/dL) | 114.82 (94.72, 136.86) | 113.66 (93.56, 134.92) | 115.59 (95.49, 138.02) | 113.85 (94.33, 138.40) | 0.11 |
| TG (mg/dL) | 106.20 (75.22, 153.99) | 107.08 (76.11, 156.11) | 106.02 (74.34, 151.34) | 106.20 (76.11, 155.76) | 0.35 |
Abbreviations: CKM Cardiovascular-Kidney-Metabolic syndrome, SBP systolic blood pressure, DBP diastolic blood pressure, TC total cholesterol, TG triglyceride, HDL high density lipoprotein, LDL Low Density Lipoprotein
Association between depressive symptom trajectories and incidence of CVD
During the follow-up period, 1,080 incident CVD events occurred. The cumulative incidence of CVD differed significantly across the depressive symptom trajectory groups: 13.0% (Low-stable), 17.3% (Moderate-stable), and 23.1% (Persistently-high), as shown in (Fig. 3). Multicollinearity analysis revealed that the VIFs for all variables in the model were less than 5, confirming no significant redundancy (Table S4). Consequently, unadjusted Cox proportional hazards models (Model 1) revealed a significantly elevated risk of incident CVD relative to the Low-stable group for both the Moderate-stable (HR = 1.35, 95% CI: 1.16–1.59; P < 0.001) and Persistently-high (HR = 1.90, 95% CI: 1.60–2.24; P < 0.001) trajectories. These associations persisted in the fully adjusted model (Model 3): Moderate-stable (HR = 1.36, 95% CI: 1.16–1.60; P < 0.001) and Persistently-high (HR = 1.88, 95% CI: 1.57–2.24; P < 0.001). Consistent risk elevations were observed for both trajectories in separate analyses of coronary heart disease and stroke (Table 2).
Fig. 3.
Cumulative incidence of CVD among participants across three depressive-symptom trajectories. Note: Bar heights represent the cumulative incidence (%) of newly onset CVD during the follow-up period across the three distinct trajectory groups. Pairwise comparisons between trajectory groups were performed using Pearson’s Chi-square tests*** P < 0.001 for pairwise comparisons
Table 2.
The relationship between the depressive symptoms trajectories and the incidence of CVD in a population with CKM stages 0–3
| Outcome | Model 1 | Model 2 | Model 3 | |||
|---|---|---|---|---|---|---|
| HR (95%CI) | P | HR (95%CI) | P | HR (95%CI) | P | |
| CVD | ||||||
| Low-stable depressive trajectory | Ref | Ref | Ref | |||
| Moderate-stable depressive trajectory | 1.35(1.16,1.59) | < 0.001 | 1.37(1.17,1.61) | < 0.001 | 1.36(1.16,1.60) | < 0.001 |
| Persistently-high depressive trajectory | 1.90(1.60,2.24) | < 0.001 | 1.91(1.60,2.28) | < 0.001 | 1.88(1.57,2.24) | < 0.001 |
| p for trend | < 0.001 | < 0.001 | < 0.001 | |||
| Heart Disease | ||||||
| Low-stable depressive trajectory | Ref | Ref | Ref | |||
| Moderate-stable depressive trajectory | 1.39(1.15,1.68) | < 0.001 | 1.39(1.14,1.68) | 0.001 | 1.38(1.13,1.67) | 0.001 |
| Persistently-high depressive trajectory | 1.83(1.49,2.25) | < 0.001 | 1.79(1.45,2.22) | < 0.001 | 1.76(1.42,2.18) | < 0.001 |
| p for trend | < 0.001 | < 0.001 | < 0.001 | |||
| Stroke | ||||||
| Low-stable depressive trajectory | Ref | Ref | Ref | |||
| Moderate-stable depressive trajectory | 1.32(1.03,1.70) | 0.03 | 1.36(1.06,1.76) | 0.02 | 1.36(1.05,1.75) | 0.02 |
| Persistently-high depressive trajectory | 2.06(1.59,2.68) | < 0.001 | 2.17(1.65,2.85) | < 0.001 | 2.14(1.63,2.82) | < 0.001 |
| p for trend | < 0.001 | < 0.001 | < 0.001 | |||
Hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated using Cox proportional hazards regression models
Subgroup and sensitivity analyses
Figure 4 and Figure S2 illustrate the associations of depressive symptom trajectories with the incidence of overall CVD, heart disease, and stroke across predefined subgroups. For the primary outcome, subgroup analyses generally demonstrated consistent associations across most strata. However, a significant interaction was observed for baseline dyslipidemia. In sensitivity analyses utilizing a complete-case approach to exclude participants with missing values (Table S5A), the associations between depressive symptom trajectories and the risks of overall CVD, coronary heart disease, and stroke remained largely consistent with the primary findings. Furthermore, the results obtained from discrete-time survival models (Table S5B), which were employed to account for the interval-censored nature of the data, were also fundamentally consistent with those derived from the primary Cox proportional hazards models.
Fig. 4.
Subgroup analysis of the associations between depressive-symptom trajectories and the incidence of CVD. Multivariable Cox proportional hazards regression was used to estimate the association between depressive-symptom trajectories and the incidence of CVD within each subgroup. Hazard ratios (HRs) with 95% confidence intervals (CIs) are presented. Interaction p-values were calculated to assess whether the effect of depressive trajectories on incident events differed across subgroups
Discussion
This study employed GBTM to identify three distinct depressive symptom trajectories among Chinese middle-aged and older adults with stage 0–3 CKM. Compared to the Low-stable trajectory group, individuals with Moderate-stable or Persistently-high trajectories exhibited significantly elevated risks of incident stroke, cardiac events, and overall CVD. These associations remained robust across subgroup and sensitivity analyses, providing novel evidence supporting the integration of longitudinal depression monitoring into cardiovascular risk assessment during early-stage CKM management.
CKM is a recently conceptualized systemic disorder characterized by complex interactions among metabolic risk factors, CKD, and CVD, which ultimately lead to increased cardiovascular event rates [20, 21]. Disease progression is modulated by a confluence of factors, including behavioral patterns [22], environmental exposures [23], and social determinants of health [24]. Although the precise mechanisms underlying the association between depression and the accelerated progression of cardiometabolic-renal diseases remain incompletely understood, converging evidence implicates several interconnected biological pathways. First, depressive states are associated with sustained hyperactivation of the sympathetic-adrenal-medullary axis, leading to chronically elevated levels of catecholamines and cortisol [25]. This neuroendocrine environment results in reductions in heart-rate variability and blood-pressure variability, both of which reflect autonomic dysregulation and ultimately promote the onset of cardiovascular events [26, 27]. Second, the hypothalamic-pituitary-adrenal (HPA) axis, which is central to stress integration, shows impaired glucocorticoid negative feedback and disrupted circadian rhythmicity in depression [25]. Its excessive activation contributes to visceral adiposity, insulin resistance, and dyslipidemia, accelerating the development of metabolic syndrome and increasing cardiovascular risk [28]. These specific neuroendocrine disruptions resonate with the recently proposed concept of Circadian Syndrome (CircS), which highlights circadian rhythm disruption as a crucial link connecting depression, short sleep, and metabolic abnormalities to incident CVD [29, 30]. Given that the CKM definition primarily focuses on disease staging rather than a common etiological origin, the CircS framework offers a critical etiological insight: depressive symptoms may not be simple comorbidities. Instead, they could act as a core pathophysiological driver that propels the progression from initial CKM stages toward severe cardiovascular complications. Previous studies have shown that higher levels of depression severity are independently associated with the progression of CKM and an increased risk of death, a conclusion validated through data analysis from two large, nationally representative cohorts. For each 1-point increase in the depression score, the risk of advanced CKM rises by 7%. In predicting the mortality rate of CKM patients, the HR was 1.04 for each 1-point increase in PHQ (1.02–1.05), and 1.02 (1.00–1.03) for every 10-point increase in CES-D [11]. In a separate cohort study, increasing baseline depressive symptom scores were associated with elevated cardiovascular risk among middle-aged and elderly individuals with CKM stages 0–3. This relationship progressively strengthened as CKM stage advanced, suggesting a synergistic interaction between mood disorders and disease progression [31].
Previous studies, as mentioned earlier, have often been limited to evaluating CKM patients based on baseline depressive symptoms, which may not fully capture the complexity of depression progression. Given population heterogeneity in depression trajectories, single-timepoint assessments may inadequately characterize individual risk profiles. Our longitudinal analysis of triennial CES-D-10 measurements revealed that Moderate-stable or Persistently-high depressive trajectories were significantly associated with a higher CVD risk compared to the Low-stable group. Specifically, in the fully adjusted model, the Persistently-high trajectory was associated with an 88% higher CVD risk, a 76% increase in heart disease risk, and a 114% elevated stroke risk. These associations persisted after comprehensive adjustment for sociodemographic, lifestyle, and clinical covariates, confirming the robust prognostic value of longitudinal depressive symptom patterns. Subgroup analyses confirmed robust depression-CVD associations across most strata. However, for overall CVD, the prognostic impact of persistently high depression was significantly stronger in individuals without baseline dyslipidemia. Pre-existing dyslipidemia may exert a ‘ceiling effect’ masking psychological distress, whereas normal lipid profiles unmask depression-triggered inflammatory vascular injury. Depression acts as an independent and ubiquitous risk multiplier across the entire CKM continuum. These findings reinforce the clinical imperative for dynamic depression surveillance and early intervention in both CKM primary prevention and secondary management.
Our study has several advantages. First, the nationally representative sampling and high response rate of CHARLS enhance the generalizability of the results to the target population. Additionally, by using repeated CES-D measurements for trajectory modeling, this approach more accurately captures the longitudinal changes in depression compared to cross-sectional assessments. However, this study also has some limitations. First, relying on self-reported physician diagnoses and CES-D scores inevitably introduces potential recall and misclassification biases. Second, excluding individuals with pre-existing cardiovascular conditions or missing data inherently introduces selection bias, which, combined with the cohort’s restriction to middle-aged and older Chinese adults, may temper the broader generalizability of our findings. Third, the biennial survey design yields interval-censored data without exact event dates, affecting time-to-event precision. Fourth, the Framingham Risk Score used for CKM staging lacks specific calibration for the Chinese population, potentially overestimating absolute cardiovascular risk. Finally, despite comprehensive covariate adjustment, residual confounding from unmeasured variables (e.g., antidepressant use) or excluded factors with high missingness (e.g., physical activity) may remain, and the observational design strictly precludes definitive causal inferences.
Conclusion
Depressive symptom trajectories show significant heterogeneity among Chinese middle-aged and older adults with CKM stages 0–3. Elevated depressive symptom trajectories, particularly the Persistently-high and Moderate-stable patterns, are independently associated with an increased risk of incident CVD. Integrating long-term tracking of depressive symptoms into routine cardiovascular risk assessments for CKM patients could help optimize prevention strategies.
Supplementary Information
Acknowledgements
None.
Authors’ contributions
M.Y. and Y.L. defined the research objectives and devised the analysis plan. They also took charge of extracting the data from CHARLS. X.P. and Y.C. played a key role in data processing and statistical analysis. M.Y. and Y.L. worked together to draft the initial manuscript. J.Y. and W.H. made significant revisions and gave their approval for the final version. All authors reviewed and approved the final manuscript.
Funding
This work was supported by the Zhejiang Provincial Public Welfare Research Project (Grant No. LGF21H250003), Zhejiang Province Traditional Chinese Medicine Science and Technology Project (Grant No. GZY-ZJ-KJ-24002), and the Zhejiang Provincial Clinical Research Center for Critical Care Medicine (2021E50005).
Data availability
The datasets analyzed in this study are publicly available through the CHARLS repository at http://charls.pku.edu.cn.
Declarations
Ethics approval and consent to participate
The CHARLS study received approval from the Institutional Review Board (IRB) of Peking University(IRB approval numbers for the household survey: IRB00001052-11015; for blood sample collection: IRB00001052-11014). Informed consent was obtained in writing from all participants.
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.
Mingkun Yang, Yanwen Lu and Yijing Liu contributed equally to this work.
Contributor Information
Weihang Hu, Email: huweihang2025@163.com.
Jing Yan, Email: yanjing201801@163.com.
References
- 1.Tsai M, Kao JT, Wong C, et al. Cardiovascular-kidney-metabolic syndrome and all-cause and cardiovascular mortality: A retrospective cohort study. PLoS Med. 2025;22(6):e1004629. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Ndumele CE, Rangaswami J, Chow SL, et al. Cardiovascular-Kidney-Metabolic Health: A Presidential Advisory From the American Heart Association. Circulation. 2023;148(20):1606–35. [DOI] [PubMed] [Google Scholar]
- 3.Ndumele CE, Neeland IJ, Tuttle KR, et al. A Synopsis of the Evidence for the Science and Clinical Management of Cardiovascular-Kidney-Metabolic (CKM) Syndrome: A Scientific Statement From the American Heart Association. Circulation. 2023;148(20):1636–64. [DOI] [PubMed] [Google Scholar]
- 4.Liu K, Hu J, Huang Y, et al. Triglyceride-glucose-related indices and risk of cardiovascular disease and mortality in individuals with cardiovascular-kidney-metabolic (CKM) syndrome stages 0–3: a prospective cohort study of 282,920 participants in the UK Biobank. Cardiovasc Diabetol. 2025;24(1):277. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Pan Y, Du B, Feng L, et al. Comparative analysis of 16 baseline obesity and lipid-related indices for cardiovascular disease risk prediction in adults with cardiovascular-kidney-metabolic syndrome stages 0–3: a nationwide prospective cohort study. Diabetol Metab Syndr. 2025;17(1):343. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.He Q, Hu F, Wei W, et al. Association between the composite dietary antioxidant index and cardiovascular-kidney-metabolic syndrome among U.S. adults: evidence from NHANES 2007–2018. Front Nutr. 2025;12:1600651. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Tu D, Zuo X, Li P, et al. Association between the life’s essential 8 health behaviors score and all-cause mortality in cardiovascular-kidney-metabolic syndrome patients. Front Nutr. 2025;12:1612693. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Dogan-Sander E, Mergl R, Willenberg A et al. Inflammation and the association of vitamin D and depressive symptomatology. Nutrients, 2021,13(6). [DOI] [PMC free article] [PubMed]
- 9.Han T, Zhang L, Jiang W, et al. Persistent depressive symptoms and the changes in serum cystatin c levels in the elderly: a longitudinal cohort study. Front Psychiatry. 2022;13:917082. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Wang G, Wu Y, Chen A, et al. Association between depressive symptoms and mortality in patients with cardiovascular-kidney-metabolic syndrome: the mediating role of inflammatory biomarkers. J Affect Disord. 2025;386:119429. [DOI] [PubMed] [Google Scholar]
- 11.He Y, Lan L, Liu Y, et al. Depression severity as a predictor of cardiovascular-kidney-metabolic syndrome progression and mortality: results from two nationally representative cohort studies. J Affect Disord. 2025;388:119606. [DOI] [PubMed] [Google Scholar]
- 12.Chen P, Jiang P, Cui Y, et al. Developmental trajectories of depressive symptoms during early adolescence: A 12-month cohort study in Nanchong, China. J Affect Disord. 2025;391:119941. [DOI] [PubMed] [Google Scholar]
- 13.Fang J, Wu W, Yang C, et al. Trajectories of depressive symptoms in middle-aged and older Chinese adults: identifying subgroups, core symptoms and predictors. BMC Psychiatry. 2025;25(1):700. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.He Y, Cao Y, Xiang R, et al. Predictive value and robustness of the stress hyperglycemia ratio combined with hypertension for stroke risk: evidence from the CHARLS cohort. Cardiovasc Diabetol. 2025;24(1):336. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Wang X, Yin Z, Gao Q, et al. Associations of indoor musty odors with depression and anxiety symptoms in Chinese older adults: a nationwide study. BMC Public Health. 2025;25(1):2793. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Shi S, Xue F, Jiang T, et al. Association between trajectory of triglyceride-glucose index and all-cause mortality in critically ill patients with atrial fibrillation: a retrospective cohort study. Cardiovasc Diabetol. 2025;24(1):278. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Huang S, Li X, Chen B, et al. Association between serum sodium trajectory and mortality in patients with acute kidney injury: a retrospective cohort study. BMC Nephrol. 2024;25(1):152. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Chen Y, Tu H, Liang M, et al. Association between body roundness index and advanced cardiovascular-kidney-metabolic syndrome. Front Nutr. 2025;12:1623766. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Liu J, Song Y, Luo R, et al. Association between urinary heavy metals and cardiovascular-kidney-metabolic syndrome: mediating roles of TyG, WWI, and eGFR. Front Nutr. 2025;12:1613721. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Ge J, Zhu L, Jiang S, et al. Association of dietary quality, biological aging, progression and mortality of cardiovascular-kidney-metabolic syndrome: insights from mediation and machine learning approaches. Nutr J. 2025;24(1):105. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Mo D, Liu T, Bai J, et al. Association between the cardiometabolic index and cardiovascular disease risk in patients with cardiovascular-kidney-metabolic syndrome: a cohort study. BMC Cardiovasc Disord. 2025;25(1):451. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Song F, Hu M, Xia J et al. Intrinsic capacity and its change predict cardiovascular disease risk in early cardiovascular-kidney-metabolic syndrome: a nationwide cohort study. Open Heart, 2025,12(2). [DOI] [PMC free article] [PubMed]
- 23.Zheng Z, Xu Y, Kang N, et al. Assessing the effect of perfluoroalkyl and polyfluoroalkyl substances on cardiovascular-kidney-metabolic syndrome: Insights from an interpretable machine learning model. Sci Total Environ. 2025;993:180003. [DOI] [PubMed] [Google Scholar]
- 24.Yang Q, Shi P, Pan L, et al. Social determinants of health (SDOH) associated with the risk of all-cause mortality and life expectancy in US adults with cardiovascular-kidney-metabolic syndrome: a NHANES 2001–2018 cohort study. BMC Cardiovasc Disord. 2025;25(1):369. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Penninx BWJH. Depression and cardiovascular disease: epidemiological evidence on their linking mechanisms. Neurosci Biobehav Rev. 2017;74(Pt B):277–86. [DOI] [PubMed] [Google Scholar]
- 26.Thayer JF, Yamamoto SS, Brosschot JF. The relationship of autonomic imbalance, heart rate variability and cardiovascular disease risk factors. Int J Cardiol. 2010;141(2):122–31. [DOI] [PubMed] [Google Scholar]
- 27.Lee E, Anselmo M, Tahsin CT, et al. Vasomotor symptoms of menopause, autonomic dysfunction, and cardiovascular disease. Am J Physiol Heart Circ Physiol. 2022;323(6):H1270–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Janssen JAMJ. New insights into the role of insulin and hypothalamic-pituitary-adrenal (HPA) axis in the metabolic syndrome. Int J Mol Sci, 2022,23(15). [DOI] [PMC free article] [PubMed]
- 29.Shi Z, Tuomilehto J, Kronfeld-Schor N, et al. The circadian syndrome predicts cardiovascular disease better than metabolic syndrome in Chinese adults. J Intern Med. 2021;289(6):851–60. [DOI] [PubMed] [Google Scholar]
- 30.Shi Z, Tuomilehto J, Kronfeld-Schor N et al. The circadian syndrome is a significant and stronger predictor for cardiovascular disease than the metabolic syndrome-The NHANES survey during 2005–2016. Nutrients, 2022,14(24). [DOI] [PMC free article] [PubMed]
- 31.Ma C, Zhao D, Zhai L, et al. Role of depressive symptoms in accelerating cardiovascular diseases progression in cardiovascular-kidney-metabolic syndrome stages 0–3: a nationwide prospective cohort study. Gen Hosp Psychiatry. 2025;96:183–92. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The datasets analyzed in this study are publicly available through the CHARLS repository at http://charls.pku.edu.cn.