The NHANES is a series of ongoing cross-sectional surveys conducted by the National Center for Health Statistics at the Centers for Disease Control and Prevention. It provides nationally representative cross-sectional data for the health status of the U.S. civilian noninstitutionalized population. Participants were interviewed at home regarding demographic, socioeconomic, dietary, and health-related questions. Subsequently, they underwent medical, dental, and physiological measurements as well as laboratory tests at a mobile examination center. Ethical approval for the study was obtained from the Research Ethics Review Board of National Center for Health Statistics, and all participants gave written informed consent. This study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.

The continuous NHANES began in 1999 and collects data biennially. In CKM syndrome stage 0 and 1, lower BMI and waist circumference cut point is advised for Asian populations given their predisposition to developing metabolic abnormalities at lower levels of adiposity (15). Since NHANES collected responses on Asian race as a separate category from the 2011–2012 cycle, we used data from four NHANES cycles, including NHANES 2011–2012, NHANES 2013–2014, NHANES 2015–2016, and NHANES 2017–2018. The link to the dataset can be found in the “Data availability” section. A total of 22,617 adults aged 20 years or older were included. We further excluded participants having the following conditions: (1) a pregnancy at baseline (n = 248); (2) ineligible data on death or follow-up (n = 1,041); (3) missing LE8 CVH scores (n = 4,431); (4) missing information on CKM syndrome indicators (n = 1,422); (5) missing information on potential covariates (n = 2,094). Consequently, the final analytical sample included in this study consisted of 13,381 adults (Figure 1).

The LE8 score comprises four health behaviors (diet, physical activity, nicotine exposure, and sleep duration) and four health factors (BMI, non-HDL cholesterol, blood glucose, and blood pressure), each with its own scoring standards ranging from 0 to 100 (10). Dietary intake was obtained by two non-consecutive 24-h dietary recalls, and the dietary quality was assessed by Healthy Eating Index 2015 (HEI-2015) scores, which aligns with the 2020–2025 Dietary Guidelines for Americans (16). Information on physical activity (self-reported minutes per week of moderate-to-vigorous physical activity), nicotine exposure (combustible cigarette use, inhaled nicotine delivery systems use, and secondhand smoke exposure) sleeping information (sleep duration), and medication use were obtained from standardized questionnaires. Height, weight, and systolic and diastolic blood pressure were measured at the physical examination. BMI was calculated as weight in kilograms divided by height in meters squared. Blood lipid and glycemic profiles were measured at a morning examination session after fasting for 9 h or more. The non-HDL cholesterol was calculated by total cholesterol minus high-density lipoprotein (HDL) cholesterol, and Hemoglobin A1c (HbA1c) was utilized to evaluate blood glucose levels. Detailed algorithms for calculating the LE8 scores for each of the eight metrics to NHANES data are provided in Supplementary Table S1. In brief, the overall LE8 score was calculated as the average of the eight CVH metrics. The overall CVH score was further categorized into low (0–49 points), moderate (50–79 points), and high (80–100 points) CVH groups, following the recommendations of AHA (10). Additionally, we used the same criteria and cut-off points to measure and categorize LE8 health behaviors and factors score to further investigate the association between LE8 subscales and all-cause mortality in patients with CKM syndrome. These definitions are consistent with previous studies (17, 18).

The AHA criteria define CKM syndrome stages based on ten components: BMI, waist circumference, hypertension, prediabetes, diabetes, hypertriglyceridemia, metabolic syndrome, CKD, 10-year CVD risk, and clinical CVD (7). Hypertension was characterized by systolic blood pressure ≥ 140 mmHg, diastolic blood pressure≥90 mmHg, a self-reported history of high blood pressure, or antihypertensive drug intake. Prediabetes was defined as having a fasting blood glucose level from 100 to 124 mg/dL, or a HbA1c of 5.7 to 6.4%. Diabetes was identified by a fasting blood glucose ≥ 125 mg/dL, a HbA1c ≥ 6.5%, self-reported history of diabetes, or current hypoglycemic treatment. Metabolic syndrome was defined as presence of any 3 of the following 5 elements: (1) the fasting blood glucose>100 mg/dL or drug treatment for diabetes mellitus; (2) HDL cholesterol <50 mg/dL in females, <40 mg/dL in males or drug treatment for reduced HDL cholesterol; (3) plasma triglyceride >150 mg/dL or drug treatment for raised triglyceride; (4) waist circumference > 88 cm in women or >102 cm in men; (5) blood pressure > 130/85 mmHg or drug treatment for raised blood pressure. Based on the Kidney Disease Improving Global Outcome criteria, CKD status was categorized as: low risk, moderate risk, high risk, and very high risk (Supplementary Table S2) (7). 10-year CVD risk was estimated with the AHA Predicting Risk of CVD EVENTs (PREVENT) equations (19).

In the present study, the definition of CKM syndrome stage was determined according to the AHA Presidential Advisory Statement on CKM syndrome (7), with modifications for NHANES data based on the criteria from Aggarwal et al. (8). Stage 0: Individuals with normal BMI and waist circumference, normoglycemia, normotension, normal lipid status, and no evidence of CKD or subclinical or clinical CVD. Stage 1: Individuals having excess weight (BMI ≥ 25 kg/m2 or ≥23 kg/m2 if Asian ethnicity), or abdominal obesity (waist circumference≥88/102 cm in women/men or 80/90 cm in women/men if Asian ethnicity), or prediabetes. Stage 2: Individuals with metabolic risk factors (hypertriglyceridemia [≥135 mg/dL], hypertension, metabolic syndrome, diabetes), or moderate-to-high-risk CKD. Stage 3: Individuals at very high risk of CKD or a high risk (≥20%) of 10-year CVD. Stage 4: Individuals with a self-reported diagnosis of CVD (coronary heart disease, angina, heart attack, heart failure, or stroke). The detailed descriptions of CKM syndrome stage definitions are in Supplementary Table S3. Based on the prevalence of CKM syndrome among U.S. adults (8), CKM syndrome included stage 1 or higher, with advanced stages being stage 3 or 4, as these indicated participants with or at high risk for CVD.

Data for deaths were obtained by linking to the NHANES-linked National Death Index until December 31, 2019 (20). Cause of death was defined using the International Statistical Classification of Disease, Tenth Revision (ICD-10). Death from all reasons was defined as all-cause mortality. Follow-up time was calculated from the date of interview to the date of death, or the end of follow-up (December 31, 2019), whichever came first.

In line with previous studies (14, 18, 21, 22), the following covariates were included the current study: age, sex (female or male), ethnicity (non-Hispanic White, non-Hispanic Black, non-Hispanic Asian, Hispanic and other race or ethnicity), education level (less than high school, high school or equivalent, and college or above), ratio of family income to poverty (less than 1, 1 to 3, and more than 3), marital status (coupled, and single or separated), BMI, HbA1c, total cholesterol, HDL cholesterol, low-density lipoprotein (LDL) cholesterol, estimated glomerular filtration rate (eGFR), alcohol consumption status (non-drinker, low to moderate drinker, and heavy drinker), history of cancer (yes or no), and the use of antihypertensive, antidiabetic, lipid-lowering medications.

All analyses were performed incorporating the sampling weights to account for the complex NHANES sampling design. For baseline characterization, continuous variables were described by weighted mean value with weighted standard errors (SE), while categorical variables were expressed as number with weighted percentage (%). To check for differences in characteristics between the three CVH groups, Analysis of Variance was used for differences in weighted means for continuous variables and the Rao-Scott Chi-Square test for differences in weighted percentages for categorical variables. Survey-weighted Cox proportional hazards models were performed to estimate hazard ratios (HR) and 95% confidence intervals (CI) for the associations of LE8 and LE8 subscales with all-cause mortality. We fitted two statistical models. Model 1 was adjusted for sociodemographic variables, including age, sex, ethnicity, educational level, family income, and marital status. Model 2 was adjusted for the variables in Model 1 and additional confounders including BMI, HbA1c, total cholesterol, HDL cholesterol, LDL cholesterol, eGFR, alcohol status, history of cancer, and the use of antihypertensive, antidiabetic, lipid-lowering medications. The Schoenfeld residuals test is used to test the proportional hazard assumption in Cox model.

To comprehensively understand the relationship between LE8 and mortality risk, restricted cubic spline (RCS) models were utilized to estimate the dose–response association of total LE8 score, LE8 health behaviors score, and LE8 health factors score with all-cause mortality. We used the Akaike information criterion (AIC) to select the RCS model with specific number of knots and adopted the model with the lowest AIC criterion value, which is considered as the best-fitting RCS model. Besides, the PAF was calculated to quantify the proportion of mortality in a population attributable to exposure to a risk factor (23). In our study, the ‘graphPAF’ R package was used to calculate the adjusted PAF of high CVH score (≥80 points) vs. moderate or low CVH score (<80 points) to quantify the proportion of all-cause mortality that could be prevented by improving CVH score.

Subgroup analyses were performed to examine the association of LE8 score with all-cause mortality stratified by age, sex, ethnicity, education level, family income, marital status, alcohol status, cancer, and the use of antihypertensive, antidiabetic, lipid-lowering medications. Sensitivity analyses were performed to assess the robustness of the results. First, we excluded participants who experienced mortality during the first year of follow-up to account for reverse causality. Second, individuals with a self-reported history of cancer were excluded. Third, iterative imputation for missing covariates was performed using a machine learning algorithm via the R package ‘missRanger’ to test the above associations. The missRanger is an iterative imputation approach based on random forest (24). Missing values for a variable are imputed using predictions made by a random forest, which employs all remaining variables as covariates. The algorithm performs repeated iterations across all variables. This process continues until there is no further improvement observed in the average Out-Of-Bag prediction error of the models, which serves as the stopping criterion for the iterations. Besides, the missRanger incorporates in the imputation process the predictive mean matching option, which is employed to avoid imputation with values not present in the original data. We employed the parameters (pmm.k = 3, num.trees = 500) in the imputation process.

All analyses were conducted using R software version 4.4.2. A two-tailed p value < 0.05 was considered statistically significant.

Of the 13,381 U.S. adults from NHANES 2011 to 2018, the proportion of those in stage 0 to 4 was 13.02, 26.84, 44.74, 4.54 and 10.86%, respectively. 86.98% of U.S. adults met criteria for CKM syndrome (stage 1 or higher) and 15.40% met criteria for advanced stages. We then removed 1,743 participants in CKM syndrome stage 0, and finally this study included 11,638 CKM syndrome patients (Figure 1).

The baseline characteristics of the CKM syndrome patients were summarized by the three categories of LE8 score in Table 1. Only 14.63% of CKM syndrome patients achieved a high LE8 score, whereas 15.04% had a low score. CKM syndrome patients with higher LE8 score were more likely to be younger, female, non-Hispanic White and Non-Hispanic Asian, coupled, non-heavy drinkers, non-cancer population, and have a higher education level, higher household income, and higher eGFR. Besides, there were lower proportions of the use of antihypertensive drugs, antidiabetic drugs, and antihyperlipidemic drugs in the higher LE8 score group. For LE8 components, participants with a high LE8 score exhibited characteristics of lower BMI, non-HDL cholesterol, HbA1c, blood pressure level, and a higher HEI-2015 score, as well as longer physical activity and sleep duration (all p < 0.0001). Besides, CKM syndrome patients in advanced stages were less likely to have high LE8 scores (all p < 0.0001).

During a median follow-up of 5.0 years, a total of 831 deaths were identified. The Schoenfeld residuals suggested that the Cox models met proportional hazards assumptions (all P for global Schoenfeld test > 0.05) (Supplementary Table S4). In the RCS analysis of total LE8 score and all-cause mortality, the use of AIC statistics indicated that the RCS function with 3 knots was the optimal model (lowest AIC: 11256.45) (Figure 2A and Supplementary Table S5). The RCS model revealed that within the context of the fully adjusted model (Model 2), there were approximately a negative linear dose–response association between increased total LE8 score and a reduced risk of all-cause mortality in CKM syndrome patients (p for non-linear > 0.05, Figure 2B). Compared to CKM syndrome patients in the low total LE8 CVH group, the moderate and high total LE8 CVH groups was associated with a 48% (HR, 0.52; 95% CI, 0.42–0.64; p < 0.0001) and 70% (HR, 0.30; 95% CI, 0.20–0.46; p < 0.0001) lower risk of all-cause mortality, respectively. HR for every 10 scores increase in total LE8 score was 0.68 (95% CI, 0.64–0.74; p < 0.0001) in association with all-cause mortality (Table 2).

When stratified by advanced stages, interestingly, a reduced risk of all-cause mortality was observed only among advanced CKM syndrome patients within the moderate and high total LE8 CVH groups when compared to all-cause mortality risk in the low total LE8 CVH group (HR, 0.51; 95% CI, 0.37–0.69; p < 0.0001 and HR, 0.35; 95% CI, 0.17–0.72; p = 0.005, respectively). However, this association was not observed among non-advanced CKM syndrome patients (HR, 0.81; 95% CI, 0.60–1.11; p = 0.19 and HR, 0.56; 95% CI, 0.27–1.16; p = 0.12, respectively). These findings suggest that the total LE8 CVH score may not be suitable for determining the risk of all-cause mortality in non-advanced CKM syndrome patients.

Given that the total LE8 CVH score contains aspects that partially overlap with the definitions of the components of CKM syndrome, such as BMI, blood lipid profile, blood glucose, and blood pressure, its effect on CKM syndrome might be lessened. Therefore, we divided the total LE8 score into two categories: health behaviors score (including diet, physical activity, nicotine exposure, and sleep duration) and health factors score (including BMI, non-HDL, blood glucose, and blood pressure). The RCS model with 3 knots (lowest AIC: 10279.18, Figure 2C and Supplementary Table S6) revealed that there were approximately a negative linear dose–response association between increased health behaviors score and a reduced risk of all-cause mortality in CKM syndrome patients (p for non-linear > 0.05, Figure 2D). Conversely, the RCS model with 6 knots (lowest AIC: 10367.34, Figure 2E and Supplementary Table S7) showed that the association of health factors score with all-cause mortality was nonlinear (p < 0.0001) and displayed a W-shaped pattern (Figure 2F). In comparison to their counterparts in the low health behaviors score group, CKM syndrome patients in the moderate and high health behaviors score groups still exhibited a significant decrease in HRs for all-cause mortality, as well as advanced and non-advanced CKM syndrome patients (Table 3). However, this significance was not observed within the health factors score category (Table 4).

Figure 3A shows that the adjusted PAF of high (≥80 points) vs. moderate or low CVH score (<80 points) with all-cause mortality in CKM syndrome patients, and the three metrics were ranked in the relative order of higher to lower fractions as follows: total LE8 score (45.8%), health behaviors score (20.0%), health factors score (11.6%). For advanced and non-advanced CKM syndrome patients, the order of PAFs remained consistent (Figures 3B,C).

Subgroup analyses by age, sex, ethnicity, education level, family income, marital status, alcohol status, cancer and the use of antihypertensive, antidiabetic, lipid-lowering medications demonstrated similar associations between LE8 health behaviors score and all-cause mortality in the CKM syndrome group (Supplementary Table S8), as well as advanced and non-advanced CKM syndrome patients (Supplementary Tables S9, S10). Besides, the patterns of moderate and high health behaviors score in relation to reduced risk of all-cause mortality remained consistent after excluding participants death within the first year of follow-up (Supplementary Table S11) or with cancer history (Supplementary Table S12), as well as iterative imputation for missing covariates (Supplementary Table S13).