In this prospective, nationwide, longitudinal cohort study involving the population aged 45 years and above in China, encompassing 3,243 participants with a 5-year follow-up, we found: (1) 776 individuals (24%) developed CMD, and 118 individuals (3.4%) developed CMM by the end of the follow-up; (2) Participants with higher baseline CVAI level and a change of elevating CVAI level may suffer an increased incidence of CMD; (3) The fully adjusted RCS regression analysis displayed a positive, linear association between cumulative CVAI and the incidence of CMD, diabetes, and stroke, but no significant association with heart disease; (4) The WQS model demonstrated a mixed effect of combined TG, age, HDL-C, WC, and BMI exposures on outcomes, and the WQS index tended to correlate positively with the risk of CMD, with TG contributing the most. In summary, our findings suggested that cumulative CVAI could be a valuable tool for the early identification of individuals at heightened risk of CMD, emphasizing the importance of long-term CVAI monitoring in clinical practice.

This study employed the CVAI, which particularly reflects Chinese visceral adiposity scores, taking into account the components of metabolic syndrome, including age, BMI, WC, TG, and HDL-C. A study conducted on a population of 12,237 Chinese individuals, with a follow-up period of 6.01 years, found that the probability of developing T2DM was 20.43% greater in the highest CVAI quartile compared to the lowest quartile [15]. In the CHARLS cohort study, Zhang et al. examined 7,070 participants and observed that there was a 57% higher risk of stroke per interquartile range increment in CVAI [31]. A population-based study of 7,439 participants aged ≥ 45 years from CHARLS with 6.01 years of follow-up found that each SD increase in CVAI was associated with a 17% increased risk for CVD, a 12% increase for heart disease, and a 31% increase for stroke [17]. A study that included 34,732 participants from the REACTION study found that CVAI was significantly associated with hypertension and prehypertension in both men and women, even after adjusting for various biochemical and lifestyle risk factors [32]. A study analyzed 42,165 Chinese individuals over a median follow-up of 3.36 years discovered that the risk of CAD was considerably more significant in the fourth CVAI quartile than in the first [23]. An investigation involving 9,280 participants from Guizhou province found that women with high CVAI had a higher chance of having a stroke. However, this relationship was not observed in male participants [33]. In addition, research conducted on individuals from a community-dwelling population in Suzhou found no significant association between CVAI and the risk of incident stroke [34]. The disparities in these studies may arise from various factors: Initially, certain studies selected participants exclusively from one province instead of using a sample that represents the entire nation, leading to selection bias. Furthermore, certain studies exclusively focused on rural people, wherein disparities in income, quality of life, education, and medical accessibility between urban and rural regions could exert a substantial impact on the outcomes. In addition, the majority of studies based CVAI on single assessments, ignoring temporal changes, potentially causing regression dilution bias and affecting result accuracy. CVAI is calculated using TG and HDL-C, which show dynamic changes. Therefore, baseline CVAI evaluation often fails to adequately capture the intricate and enduring alterations linked to the advancement of the disease, which are essential for prognostic assessment and clinical application.

Unlike single baseline CVAI measurements, our study utilized cumulative CVAI to evaluate the relationship between CVAI exposure and incident CMD, leveraging long-term follow-up data. Compared to single baseline CVAI, cumulative CVAI had a more significant impact on outcomes, offering a more reliable assessment method, and yielding more robust and stable results. In this study, we found that cumulative CVAI was associated with an increased risk of developing CMD. The risk of CMD development was highest in individuals in the highest quartile group, with a multivariate-adjusted OR of 1.63. Additionally, this risk was not attenuated by additional adjustment for baseline CVAI. The cumulative effect appears to be independent of and superior to baseline CVAI in the pathogenesis of CMD. Likewise, the longest exposure time was associated with the highest risk of CMD. These findings suggested that cumulative CVAI monitoring has substantial clinical potential, allowing for a more comprehensive evaluation of cardiovascular and metabolic changes in participants during follow-up, thereby aiding clinicians in better long-term health management and risk evaluation.

Our RCS model showed that cumulative CVAI was linearly associated with CMD, diabetes, and stroke but not with heart disease. Consistent with our findings, linear associations between CVAI and stroke were observed in several prospective cohort studies [27, 29, 33]. A Mendelian randomization study also confirmed a positive association between CVAI and stroke risk, further supporting the linear relationship between CVAI and incident stroke [35]. Additionally, Pan et al. reported a nonlinear (U-shaped) relationship between CVAI and T2DM risk in the RCS model [36]. As CVAI increases, its influence on T2DM incidence tends to stabilize at higher values. This difference may be attributed to our use of cumulative CVAI, which more sensitively indicated that high cumulative CVAI significantly increased diabetes incidence. This suggested that in clinical management, early disease prediction and intervention for populations with high cumulative CVAI, rather than baseline indicators, may be more effective in reducing CMD risk. However, inconsistent with prior study [17], there was no significant association between cumulative CVAI and heart disease. These inconsistent results on the association may be attributable to the sample size, the length of follow-up, the method of statistical analysis, the adjusted covariables, environment, or other factors. Prospective studies with a longer duration of follow-up and larger sample size are needed in future for an in-depth evaluation of the association between CVAI and CVD.

Another notable point was that TG was the primary contributor to the observed association between cumulative CVAI and CMD risk from the WQS regression model. Consistent with our findings, Huo et al. elucidated the underlying mechanisms of the TyG-BMI by highlighting TG as the primary contributor to the observed association between cumulative TyG-BMI and stroke risk [37]. The potential mechanisms linking increased CMD risk with elevated TG levels may be explained by chronic inflammation, insulin resistance, and endothelial dysfunction. In addition, age also identified an essential contributor to the observed association between cumulative CVAI and CMD risk, which suggested the importance of monitoring long-term CVAI changes in the middle-aged and elderly population.

The exact physiological mechanisms explaining the association between CVAI and CMD remain unclear, but several hypotheses have been proposed. First, visceral adiposity induces a systemic inflammatory state, particularly evident in vascular inflammation, by increasing the expression of interleukin-6, tumor necrosis factor-a, and high-sensitivity C-reactive protein, leading to CMD [12, 38]. Second, visceral obesity exacerbates the production of inflammatory markers and adipocytokines while reducing the production of adiponectin, leading to IR and consequently increasing the incidence of CMD [39, 40]. Third, visceral adiposity could lead to renal cytokine imbalance and damage to the glomerular basement membrane, initiating metabolic dysfunction in the kidneys [41]. Then excessive reactive oxygen species and reactive nitrogen species were produced, which induced oxidative stress, presenting as oxidized low-density lipoprotein, 8-hydroxylated deoxyguanosine, malondialdehyde, thioredoxin, and advanced oxidation protein products [42]. Oxidative stress induces a vicious cycle of endothelial dysfunction, inflammation, and fibroblast proliferation and affects arteries through stenosis and occlusion, leading to CMD incidence [39]. Last, in addition to inherited genetic factors, we found that individuals with high cumulative CVAI had high SBP, DBP, TC, TG, GLU, Cr, UA, BMI, and WC values, and the majority had hypertension or dyslipidemia, which were primary CMD risk factors. However, biologically meaningful effects might not have been completely eliminated, even with adjustment for the above confounders. Consequently, the risk of CMD was obviously elevated under high cumulative CVAI exposure, which could have been due to the synergistic effects of these factors.

The present findings have significant value for the prevention and management of CMD among the middle-aged and elderly Chinese population. Currently, the CVAI is widely used clinically as a surrogate indicator of visceral obesity, further refining the assessment of CMD risk. More attention should be given to the long-term hazards associated with the cumulative exposure and long duration of high CVAI values rather than focusing on only a single CVAI measurement during routine clinical evaluation. Using data from electronic medical records, CVAI values were automatically generated from traditional indicators, and we utilized repeated measurement data at different time points to capture the dynamic cumulative changes. Thus, measures for implementing electronic medical record information management and popularizing personal dynamic monitoring devices will provide future directions for the primary prevention of CMD. Importantly, considering that the lifetime CMD risk depends on early cumulative exposure to risk factors, identification of high-risk individuals and timely intervention to reduce cardiovascular symptoms and events have relevant practical implications.

The strengths of this study include: (1) The current study is comprised of the abundant, credible medical data of CHARLS, and the prospective nationwide cohort has extended follow-up. Moreover, the CHARLS study adopted a standardized protocol for multiple potential confounders, including anthropometric measurements, lifestyle behaviors, and laboratory indicators, to ensure the quality of data collection; (2) The use of a scientific machine learning method (K-means clustering) to explore the relationship between changes in CVAI and CMD, instead of baseline CVAI, which could represent the longstanding status of IR. To the best of our knowledge, this was the first study to undertake such an assessment; (3) We adjusted for underlying confounding factors in analyses where possible and conducted subgroup and sensitivity analyses to control for bias to guarantee the robustness of the results; 4) We have provided new evidence for the primary prevention of CMD, with an expectation of lowering the incident rate of CMD via early recognition and intervention in populations with high CVAI exposure.

However, this study also had limitations: (1) Akin to challenges encountered in similar studies, some CMD diagnoses were based on self-reports from participants, which may underestimate the actual prevalence and also cannot distinguish specific types of heart disease. The CHARLS lacked medical records, precluding the validation of these self-reported CMD cases, highlighting a gap that future large-scale, randomized controlled trials could aim to fill; (2) Despite efforts to encompass a wide range of potential confounders, other relevant factors, such as environmental changes and genetic susceptibility, were not considered owing to the limitations of the study design; (3) Due to the expansive scale of the cohort and budgetary constraints, MRI and CT scans—the gold standards for visceral fat assessment—were not employed to verify coherence with the actual amount of visceral fat and the cumulative CVAI values; (4) The study subjects were middle-aged and elderly Chinese individuals, and the conclusions may mainly apply to East Asian populations, with applicability to people under 45 years of age still unclear; (5)Although this study found an independent positive relationship between CVAI and CMD and CMM, it did not further explore the conversion model between CMD and CMM and the prognosis of CMM patients, which will be the focus of subsequent research.