Association Between NHHR and Future Cardiovascular Disease Among Patients With CKM Stages 0–3: A Nationwide Prospective Cohort Study

ABSTRACT

The American Heart Association recently proposed the concept of Cardiovascular‐kidney‐metabolic (CKM) syndrome, emphasizing the interconnections among cardiovascular disease (CVD), chronic kidney disease (CKD), and metabolic disorders. The ratio of non‐high‐density lipoprotein cholesterol to high‐density lipoprotein cholesterol (NHHR) has emerged as a novel lipid marker associated with CVD, but its relevance in individuals with CKM syndrome remains unclear. This study aimed to examine the association between NHHR and incident CVD among adults with CKM stages 0–3 using data from the China Health and Retirement Longitudinal Study (CHARLS). CVD events were defined as self‐reported heart disease or stroke. Cox proportional hazards models and restricted cubic spline (RCS) analyses were applied to assess associations, and receiver operating characteristic (ROC) curves evaluated predictive performance. Among 7445 participants (mean follow‐up: 80 months), 1476 developed CVD. Each one‐unit increase in NHHR was associated with a 4% higher risk of CVD (95% CI: 1.00–1.08). Participants in the highest quartile (Q4) had a 23% higher risk compared with Q1 (HR = 1.23, 95% CI: 1.03–1.47). RCS analysis showed a significant positive linear relationship (P overall = 0.018). NHHR demonstrated the strongest predictive ability for CVD, with consistent results across subgroups. Higher NHHR levels were independently linked to increased CVD risk among individuals with CKM stages 0–3, suggesting NHHR may serve as a simple marker to identify high‐risk populations.

1. Introduction

Cardiovascular‐kidney‐metabolic (CKM) syndrome is an integrative concept newly proposed by the American Heart Association (AHA) in 2023 [1]. It elucidates the complex pathophysiological interplay among cardiovascular disease (CVD), chronic kidney disease (CKD), and metabolic disorders (MetS) [2]. Studies have shown that these conditions are interconnected through shared mechanisms such as insulin resistance, chronic inflammation, neurohormonal activation, and endothelial dysfunction, forming a vicious cycle that accelerates target organ damage [3]. The global burden of CKM syndrome is increasing rapidly—approximately 90% of American adults are at some stage of CKM syndrome—and China's rapidly aging population has made CKM syndrome an emerging major public health challenge [4].

The CKM staging system (stages 0–4) provides an innovative framework for dynamic risk assessment throughout the disease continuum—from the appearance of risk factors to the onset of end‐stage organ damage [5]. Importantly, the early stages of CKM (stages 0–3) encompass a broad population ranging from individuals without risk factors to those with subclinical cardiovascular alterations, representing a crucial “window of opportunity” for CVD prevention [6]. The AHA has emphasized that research focusing on CKM stages 0–3 should prioritize primary prevention of cardiovascular events, underscoring the urgency of identifying reliable early‐stage risk predictors [1].

Among the multiple risk factors contributing to CKM syndrome, lipid metabolism disorders play a central role [7]. Traditionally, low‐density lipoprotein cholesterol (LDL‐C) has been the primary target for lipid management and CVD risk prediction [8]. However, increasing evidence indicates that even when LDL‐C levels are within target ranges, patients may still face substantial “residual cholesterol (RC) risk,” prompting researchers to explore more comprehensive lipid indicators [9]. Non‐high‐density lipoprotein cholesterol (non‐HDL‐C), which includes cholesterol carried by all atherogenic lipoproteins (e.g., LDL, VLDL, IDL), has been proposed as a superior marker of atherosclerotic risk compared with LDL‐C alone [10].

In recent years, several composite lipid indices have demonstrated strong predictive value for CVD risk, including the triglyceride–glucose index (TyG), cardiometabolic index (CMI), atherogenic index of plasma (AIP), and RC [11, 12]. Among them, the ratio of non‐HDL‐C to HDL‐C (NHHR) has emerged as a novel integrated lipid parameter, theoretically reflecting the balance between atherogenic and antiatherogenic lipoproteins [13]. Previous studies have suggested that NHHR shows excellent predictive ability for several CVD‐related risk factors such as diabetes, hypertension, and nonalcoholic fatty liver disease, even outperforming traditional single lipid indicators [14, 15]. Compared with individual lipid measures, NHHR may better capture the overall state of lipid metabolism, providing improved risk stratification capacity [16]. Despite its potential, most existing studies on NHHR have focused on specific populations, such as patients with diabetes or coronary artery disease [15]. Research examining NHHR within the broader and more clinically relevant context of CKM syndrome remains limited. Based on this background, we hypothesized that baseline NHHR levels are independently associated with the risk of incident CVD among Chinese middle‐aged and older adults with CKM stages 0–3, and that NHHR may outperform traditional lipid parameters (e.g., LDL‐C, non‐HDL‐C) in predicting CVD risk. Our findings are expected to provide scientific evidence to guide early intervention and precision prevention strategies for CKM syndrome.

To our knowledge, this is the first study to investigate NHHR within the prospective integrative framework of CKM syndrome using a nationally representative cohort of Chinese middle‐aged and older adults. The present study aims to address a critical question: Is baseline NHHR independently associated with incident CVD during the key preventive window of CKM stages 0–3? The results may offer high‐level epidemiological evidence to support proactive, population‐based cardiovascular prevention strategies.

2. Methods

2.1. Study Population

Data for this study were obtained from the China Health and Retirement Longitudinal Study (CHARLS), a nationwide, large‐scale public health survey designed to investigate the health, economic, and social conditions of middle‐aged and older adults in China [17]. The baseline survey was conducted between 2011 and 2012 across 150 counties or districts in 28 provincial‐level administrative regions, using a multistage, stratified, probability‐proportional‐to‐size sampling method. More than 17 000 households were included, ensuring strong national representativeness of adults aged ≥45 years. The study was approved by the Institutional Review Board of Peking University (IRB00001052‐11015), and all participants provided written informed consent at enrollment [18]. Since baseline, follow‐up surveys have been conducted every 2–3 years to collect extensive data on demographics, health status, healthcare utilization, and biomarkers. The present study followed the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines.

The participant selection process is illustrated in Figure 1. A total of 17 705 participants were initially enrolled. Exclusion criteria were as follows: (1) Prevalent CVD or stroke at baseline or missing related information (n = 2747); (2) Diagnosis of cancer (n = 134); (3) Missing key covariates, including body mass index (BMI, n = 3260), age (n = 464), CKM stage (n = 3204), smoking status (n = 24), blood pressure (n = 97), hypertension (n = 36), and diabetes (n = 51); and (4) Follow‐up duration less than two years (n = 241). Ultimately, 7445 participants were included in the final analysis and categorized into four quartiles based on NHHR levels (Q1–Q4, with 1862, 1862, 1862, and 1862 participants, respectively).

FIGURE 1.

Flowchart of the participants’ selection.

2.2. Definition of Exposure and Outcomes

The exposure variable in this study was the NHHR, calculated using fasting blood samples collected at baseline during the first wave (2011) of the CHARLS. Total cholesterol (TC) and HDL‐C were measured by an enzymatic colorimetric method, with coefficients of variation of 0.80% and 1.00%, respectively. Baseline NHHR was computed as follows [19, 20]:

NHHR = (TC−HDL‐C)/HDL‐C

The baseline NHHR was used as the exposure variable to evaluate its association with incident CVD events during follow‐up. The outcome of interest was the occurrence of new‐onset CVD or stroke (CVD‐new) during follow‐up from baseline through 2020. CVD was defined as a composite endpoint including myocardial infarction, coronary atherosclerotic heart disease, angina pectoris, congestive heart failure, and other physician‐diagnosed cardiac diseases [21]. Incident cases were identified based on participants’ self‐reports during follow‐up interviews using standardized CHARLS questionnaires. Specifically, participants were asked, “Have you ever been diagnosed by a doctor with a heart attack, coronary heart disease, angina, congestive heart failure, or other heart problems?” Stroke cases were identified through affirmative responses to “Have you ever been diagnosed by a doctor with a stroke?” or “Are you currently receiving any treatment (Chinese medicine, Western medicine, physical therapy, acupuncture, or occupational therapy) for stroke?” [22] To ensure data reliability, CHARLS implemented strict and standardized quality control procedures across data collection, recording, and verification processes.

2.3. Definition of CKM Stages 0–3

The classification of CKM stages was based on the AHA's 2023 framework for CKM syndrome [23]. The study population was strictly limited to individuals in CKM stages 0–3:Stage 0: Ideal cardiovascular, renal, and metabolic health, defined as optimal levels of BMI, blood glucose, blood pressure, and lipid profile, with no evidence of CKD or subclinical/clinical CVD. Stage 1: Characterized by excess or dysfunctional adiposity, defined by overweight (BMI ≥ 23 kg/m²), abdominal obesity (waist circumference ≥ 90 cm in men or ≥ 80 cm in women), or evidence of adipose dysfunction (e.g., impaired glucose tolerance), without additional metabolic risk factors, CKD, or CVD. Stage 2: Presence of metabolic risk factors (e.g., dyslipidemia, hypertension, metabolic syndrome, or diabetes) or moderate‐risk CKD. Stage 3: Defined as subclinical CVD without overt symptoms. Individuals with very high‐risk CKD were considered risk‐equivalent. eGFR was calculated using the modified Chinese MDRD (C‐MDRD) equation and staged according to KDIGO guidelines [1]. Since this study focused on early CKM stages and incident CVD risk, participants with clinically diagnosed CVD (CKM stage 4) at baseline were excluded. Detailed definitions of CKM syndrome stages 0–3 are provided in Table S1 of the Supplementary Material.

2.4. Definition of Covariates

Baseline covariates were collected in CHARLS using standardized questionnaires, physical examinations, and laboratory assessments. Demographic variables included age, sex, education level (less than primary, primary, junior high, senior high, or above), and marital status. Lifestyle factors included smoking status (never, former, current) and drinking status (yes/no). Anthropometric measurements included systolic blood pressure (SBP), diastolic blood pressure (DBP), waist circumference, and BMI.Laboratory indicators comprised lipid profiles (TC, triglycerides [TG], HDL‐C, LDL‐C), glycemic indices (fasting blood glucose [FBG], glycated hemoglobin [HbA1c]), renal function markers (blood urea nitrogen [BUN], serum creatinine [Cr], eGFR, cystatin C), uric acid (UA), C‐reactive protein (CRP), and hematologic parameters (platelet count and hemoglobin). Medical history was defined based on clinical measurements, self‐reported physician diagnoses, and medication use. Hypertension was defined as SBP ≥ 140 mmHg and/or DBP ≥ 90 mmHg, physician diagnosis, or use of antihypertensive medication [24]. Diabetes was defined as FBG ≥ 126 mg/dL, HbA1c ≥ 6.5%, physician diagnosis, or use of hypoglycemic drugs [25]. Additional comorbidities, including pulmonary disease, liver disease, cancer, and CVD, were defined by self‐reported physician diagnoses. Final CKM staging (0–3) was determined according to AHA criteria based on these parameters.

2.5. Statistical Analysis

Participants were categorized into four groups according to NHHR quartiles (Q1–Q4). Continuous variables with normal distributions were presented as mean ± standard deviation and compared using one‐way analysis of variance (ANOVA). Skewed continuous variables were expressed as median (interquartile range) and compared using the Kruskal–Wallis H test. Categorical variables were expressed as number (percentage) and compared using the χ ² test. To prospectively assess the association between NHHR and incident CVD risk, both univariate and multivariate Cox proportional hazards regression models were constructed to estimate hazard ratios (HRs) and 95% confidence intervals (CIs). Three models were developed: Model 1: Unadjusted. Model 2: Adjusted for age, gender, marital status, education level, and urban/rural residence. Model 3: Further adjusted for SBP, DBP, smoking, drinking, BMI, platelet count, BUN, and serum creatinine. Receiver operating characteristic (ROC) curve analysis was conducted to compare the discriminative ability of NHHR with other lipid indicators (LDL‐C and non–HDL‐C). The area under the curve (AUC) was calculated for each parameter. ROC curves were also generated based on the three Cox regression models to compare predictive performance. Subgroup analyses were conducted stratified by age, sex, smoking status, drinking status, and CKM stage (0–3), and likelihood ratio tests were used to assess potential interaction effects. Restricted cubic spline (RCS) functions were fitted in the Cox model to visually explore potential nonlinear associations between NHHR and CVD risk. The variance inflation factor (VIF) for all covariates was < 5, indicating no significant multicollinearity. Sensitivity analyses were performed using datasets after multiple imputation for missing data to verify the robustness of the results. All statistical analyses were performed using R software (version 4.4.0) and Empowerstats software, and a two‐sided p value < 0.05 was considered statistically significant.

3. Result

3.1. Baseline Characteristics of Participants

This study used 2011 as the baseline and initially enrolled 17 705 participants. By the end of 2018, the mean follow‐up was 80 months (SD 12.7) and the median was 84 months; 241 individuals were lost, leaving 7445 participants for the final analysis. A total of 7445 participants were included and categorized into four quartiles according to NHHR levels (Q1–Q4). Baseline characteristics are presented in Table 1. During the average follow‐up period of 80 months, 1476 participants developed CVD or stroke. Overall, 52.7% were male and 47.3% were female, with a mean age of 58.6 ± 8.7 years. With increasing NHHR levels, the proportions of current smokers and drinkers decreased. Metabolic parameters, including SBP, DBP, waist circumference, BMI, FBG, glycated hemoglobin (HbA1c), UA, TC, TG, and LDL‐C, all showed upward trends, whereas HDL‐C levels decreased progressively (p < 0.01 for all). The prevalence of hypertension, diabetes, CVD, and metabolic syndrome was significantly higher among participants with elevated NHHR (p < 0.01). The distribution of CKM stages also differed significantly across quartiles, with the highest proportion of CKM stage 3 observed in the Q4 group compared to Q1 (11.06% vs. 1.18%, p < 0.01). Additionally, to comprehensively illustrate the baseline characteristics of the enrolled population, Table S2 in the supplementary material presents these characteristics stratified by both CKM stages 0–3 and clinical outcomes.

TABLE 1.

Baseline characteristics by categories of NHHR level among individuals with CKM syndrome stages 0–3.

Characteristic	Q1(n = 1862)	Q2(n = 1862)	Q3(n = 1860)	Q4(n = 1861)	p‐value
Age, years	58.76 ± 8.90	58.68 ± 8.84	58.35 ± 8.39	58.50 ± 8.51	>0.05
Education level					0.006
No completion of primary school	939 (50.48%)	884 (47.53%)	885 (47.58%)	844 (45.35%)
Sishu/home school/elementary school	402 (21.61%)	427 (22.96%)	402 (21.59%)	409 (21.98%)
Middle school	361 (19.41%)	356 (19.14%)	405 (21.75%)	386 (20.80%)
High school and above	158 (8.49%)	194 (10.38%)	168 (9.08%)	221 (11.88%)
Married					>0.05
Yes	232 (12.46%)	202 (10.85%)	187 (10.04%)	190 (10.20%)
No	1630 (87.54%)	1660 (89.15%)	1673(89.96%)	1671 (89.80%)
Gender					<0.01
Female	901 (48.39%)	1035 (55.67%)	991 (53.22%)	994 (53.44%)
Male	961 (51.61%)	825 (44.33%)	869 (46.78%)	867 (46.56%)
Smoking statues					<0.01
Never	1060 (56.93%)	1164 (62.55%)	1145(61.49%)	1152 (61.87%)
Former	123 (6.61%)	134 (7.20%)	150 (8.06%)	184 (9.88%)
Current	679 (36.47%)	563 (30.25%)	567 (30.45%)	526 (28.25%)
Drinking statues					<0.01
No	1095 (58.81%)	1223 (65.72%)	1266(67.99%)	1301 (69.87%)
Yes	767 (41.19%)	638 (34.28%)	596 (32.01%)	561 (30.13%)
SBP, mmHg	126.69 ± 20.50	128.10 ± 21.00	130.41±20.87	133.35± 20.89	<0.01
DBP, mmHg	73.56 ± 12.03	74.59 ± 11.98	76.27 ± 11.90	77.80 ± 11.77	<0.01
Waist circumference, cm	79.32 ± 10.44	82.47 ± 11.96	85.43 ± 12.38	88.83 ± 11.88	<0.01
BMI, kg/m2	21.73 ± 3.25	22.95 ± 3.78	23.98 ± 3.65	25.01 ± 3.77	<0.01
PLT, (×10^9/L)	202.12 ± 70.75	210.34 ± 74.06	212.50±70.76	222.51± 73.70	<0.01
BUN, mg/dl	16.15 ± 4.82	15.69 ± 4.41	15.57 ± 4.36	15.47 ± 4.23	<0.01
FBG, mg/dl	102.77 ± 25.41	105.40 ± 27.70	109.73±37.28	120.28± 45.86	<0.01
Scr, mg/dL	0.77 ± 0.31	0.76 ± 0.18	0.78 ± 0.18	0.80 ± 0.20	<0.01
TC, mg/dl	174.10 ± 31.22	187.02 ± 32.41	197.19±33.26	216.19± 41.76	<0.01
TG, mg/dl	76.92 ± 32.64	100.01 ± 43.79	127.41±55.77	210.38±114.35	<0.01
HDL‐c, mg/dl	67.36 ± 14.54	54.86 ± 9.89	46.88 ± 8.30	37.20 ± 8.22	<0.01
LDL‐C, mg/dl	94.42 ± 23.33	114.44 ± 25.84	125.77±29.43	131.29± 44.34	<0.01
CRP, mg/dl	2.29 ± 7.44	2.54 ± 7.52	2.59 ± 6.52	2.96 ± 7.63	<0.05
HBA1C,%	5.11 ± 0.59	5.20 ± 0.67	5.28 ± 0.80	5.44 ± 1.04	<0.01
UA, mg/dl	4.25 ± 1.20	4.25 ± 1.15	4.44 ± 1.19	4.78 ± 1.31	<0.01
EGFR_SCR	98.06 ± 12.91	97.09 ± 13.08	96.71 ± 12.81	94.65 ± 14.97	<0.01
HCT,%	40.56 ± 6.27	41.20 ± 6.12	41.85 ± 6.28	42.49 ± 6.07	<0.01
HGB, g/dl	14.07 ± 2.19	14.18 ± 2.17	14.53 ± 2.26	14.73 ± 2.17	<0.01
Cystatin C, mg/l	1.03 ± 0.30	1.02 ± 0.27	1.00 ± 0.23	0.97 ± 0.27	<0.01
Hypertension					<0.01
No	1533 (82.33%)	1445 (77.59%)	1368(73.58%)	1228 (66.00%)
Yes	329 (17.67%)	417 (22.41%)	492 (26.42%)	633 (34.00%)
Diabetes					<0.01
No	1818 (97.64%)	1785 (95.86%)	1764(94.84%)	1705 (91.62%)
Yes	44 (2.36%)	77 (4.14%)	96 (5.16%)	156 (8.38%)
CVD					<0.01
No	1560 (83.78%)	1477 (79.31%)	1484(79.81%)	1447 (77.77%)
Yes	302 (16.22%)	385 (20.69%)	376 (20.19%)	414 (22.23%)
Lung disease					<0.01
0	1667 (89.67%)	1697 (91.08%)	1715(92.31%)	1720 (92.58%)
1	192 (10.33%)	165 (8.92%)	143 (7.69%)	138 (7.42%)
Liver disease					>0.05
0	1790 (96.44%)	1799 (96.93%)	1806(97.25%)	1813 (97.63%)
1	66 (3.56%)	57 (3.07%)	51 (2.75%)	44 (2.37%)
MetS					<0.01
0	508 (27.28%)	309 (16.60%)	151 (8.11%)	22 (1.18%)
1	1354 (72.72%)	1552 (83.40%)	1711(91.89%)	1840 (98.82%)
CKM stage					<0.01
0	476 (25.56%)	292 (15.69%)	143 (7.68%)	20 (1.07%)
1	289 (15.52%)	453 (24.40%)	576 (30.99%)	681 (36.57%)
2	1075 (57.73%)	1080 (57.93%)	1085(58.32%)	956 (51.29%)
3	22 (1.18%)	37 (1.99%)	56 (3.01%)	206 (11.06%)

Abbreviation: BMI, body mass index; BUN, blood urea nitrogen; CKM, cardiovascular‐kidney‐metabolic; CRP, c‐reactive protein; CVD, cardiovascular disease; DBP, diastolic blood pressure; eGFR, estimated glomerular filtration rate; FBG, fasting blood glucose; HbA1c, hemoglobin A1c; HCT, hematocrit; HDL‐c, high‐density lipoprotein cholesterol; HGB, hemoglobin; LDL‐C, low‐density lipoprotein cholesterol; MetS, metabolic syndrome; NHHR, highdensity lipoprotein cholesterol ratio; SBP, systolic blood pressure; Scr, serum creatinine; TC, total cholesterol; TG, triglycerides; UA, uric acid; WC, waist circumference.

3.2. Association Between NHHR and Incident CVD among Individuals With CKM Stages 0–3

As shown in Table 2, multivariable Cox proportional hazards models were applied to evaluate the association between NHHR and the risk of incident CVD. When NHHR was analyzed as a continuous variable, each one‐unit increase in NHHR was associated with a 7% higher risk of CVD in the unadjusted model (Model I) (HR = 1.07, 95% CI: 1.04–1.11, p < 0.01). After adjustment for partial confounders in Model II, this association remained significant, with a 7% increased risk per unit increment of NHHR (HR = 1.07, 95% CI: 1.04–1.11, p < 0.01). In the fully adjusted Model III, each one‐unit increase in NHHR was still associated with a 4% higher risk of incident CVD (HR = 1.04, 95% CI: 1.00–1.08, p < 0.05).

TABLE 2.

Cox proportional hazards model results for the association between NHHR levels and risk of new‐onset cardiovascular disease and stroke in individuals with CKM syndrome stages 0–3.

Exposure	Model I (HR.95%CI)	p	Model II (HR.95%CI)	p	Model III (HR.95%CI)	p
NHHR	1.07 (1.04, 1.11)	<0.01	1.07 (1.04, 1.11)	<0.01	1.04 (1.00, 1.08)	<0.05
NHHR quartile
Q1	Ref		Ref		Ref
Q2	1.35 (1.14, 1.59)	<0.01	1.31 (1.11, 1.55)	<0.01	1.24 (1.05, 1.47)	<0.05
Q3	1.31 (1.11, 1.55)	<0.01	1.30 (1.10, 1.54)	<0.01	1.16 (0.98, 1.39)	>0.05
Q4	1.48 (1.25, 1.74)	<0.01	1.44 (1.22, 1.71)	<0.01	1.23 (1.03, 1.47)	<0.05

Model 1: Non‐adjusted.

Model 2: Adjusted for age, gender, education level, marry, rural.

Model 3: Adjusted for age, gender, education level, marry, rural, alcohol intake, smoking, SBP, DBP, PLT, BUN, Scr, BMI.

Abbreviation: BMI, body mass index; BUN, blood urea nitrogen; CI, confidence interval; CKM, cardiovascular‐kidney‐metabolic; CVD, cardiovascular disease; HR, hazard ratio; NHHR, highdensity lipoprotein cholesterol ratio; PLT, platelet count; Scr, serum creatinine.

To further clarify the relationship between NHHR and CVD incidence, NHHR was divided into quartiles. Compared with participants in the lowest quartile (Q1), those in Q2, Q3, and Q4 showed progressively elevated risks of developing CVD in the fully adjusted Model III, with corresponding HRs of 1.24 (95% CI: 1.05–1.47), 1.16 (95% CI: 0.98–1.39), and 1.23 (95% CI: 1.03–1.47), respectively. Except for Q3, where the association was not statistically significant, higher NHHR levels were significantly associated with an increased risk of incident CVD (p < 0.05).

As shown in Figure 2, using the RCS Cox proportional hazards regression model, we observed a significant association between NHHR and the risk of incident CVD (p for overall association = 0.018). The relationship exhibited a monotonically increasing trend, indicating that higher NHHR levels were associated with a progressively greater risk of developing CVD.

FIGURE 2.

The RCS analysis between the NHHR level and CVD incidence in a population with CKM syndrome stages 0–3.

Although the curve suggested a slightly accelerated increase in risk at higher NHHR levels, the test for nonlinearity yielded a p value of 0.094 (> 0.05), indicating no statistically significant nonlinearity. These findings therefore support a positive linear association between NHHR and CVD risk.

The model was adjusted for Age, Gender, Smoking statues, Drinking statues, Education level, Marital status, BUN, Scr, SBP, DBP TC, CRP, Hypertension, Diabetes.

3.3. Predictive Performance of NHHR, LDL‐C, and Non–HDL‐C for CVD Risk in Patients With CKM Syndrome

As illustrated in Figure 3, ROC curve analysis was conducted to evaluate the predictive performance of three lipid‐related indices for incident CVD. The AUC values were 0.594 for NHHR, 0.534 for LDL‐C, and 0.581 for non–HDL‐C, respectively. Among individuals with CKM stages 0–3, NHHR demonstrated the highest discriminative ability for predicting future CVD events, outperforming both LDL‐C and non–HDL‐C. This finding suggests that NHHR may serve as a more sensitive early marker for cardiovascular risk, potentially facilitating earlier clinical recognition and the implementation of proactive preventive and therapeutic strategies.

FIGURE 3.

Comparative discrimination performance of different lipid markers in predicting new‐onset CVD.

To further explore the relationship between NHHR and the incidence of CVD, subgroup and interaction analyses were performed stratified by age group, sex, smoking status, drinking status, and CKM stage (0–3). As shown in Figure 4, no significant interaction effects were observed across any subgroups (P for interaction > 0.05). This indicates that the association between NHHR and CVD risk was consistent and robust among different demographic and clinical strata.

FIGURE 4.

Subgroup analyses of the association between NHHR and incident cardiovascular disease among participants with CKM stages 0–3: The model was adjusted for age, gender, smoking statues, drinking statues, education level, marital status, BUN, Scr, SBP, DBP TC, CRP, hypertension, diabetes.

4. Discussion

We found that higher NHHR levels were significantly associated with an increased risk of CVD, and this relationship remained robust even after full adjustment for potential confounders. Furthermore, RCS analysis demonstrated a significant linear dose–response association between NHHR and CVD risk across CKM stages 0–3. These findings suggest that NHHR may serve as a reliable biomarker for predicting CVD risk in early‐stage CKM populations.

As a composite lipid parameter, NHHR provides a more comprehensive reflection of lipid metabolism by simultaneously quantifying the balance between atherogenic and antiatherogenic lipoproteins [26]. Non–HDL‐C encompasses all cholesterol contained in atherogenic lipoproteins—including LDL, VLDL, IDL, and lipoprotein(a)—and is therefore considered a more complete marker of “bad cholesterol” than LDL‐C alone [27]. Clinical guidelines recommend emphasizing non–HDL‐C, particularly in patients with diabetes, obesity, or metabolic syndrome, who often present with elevated non–HDL‐C and TG and reduced HDL‐C, even when LDL‐C levels are within the normal range [28]. HDL‐C, in contrast, plays a protective vascular role by promoting reverse cholesterol transport and exerting anti‐inflammatory and antioxidant effects [29]. Within the CKM pathophysiological framework, insulin resistance and chronic inflammation synergistically drive typical lipid alterations—elevated triglyceride‐rich lipoproteins and dysfunctional HDL [30]. Thus, elevated NHHR accurately captures the lipid imbalance central to CKM syndrome, reflecting the enhancement of atherogenic forces and the weakening of protective mechanisms. This may explain why NHHR exhibited superior predictive capacity in our cohort. Especially importantly, within the CKM syndrome framework that integrates cardiovascular, renal, and metabolic risks, the lipid disorder reflected by an elevated NHHR is likely not only a driver of atherosclerosis but also linked to the onset and progression of CKD through shared pathophysiological pathways. Insulin resistance, chronic low‐grade inflammation, and oxidative stress constitute common soil connecting CKD with CVD [31]. Elevated non–HDL‐C can directly contribute to renal injury by inducing damage to glomerular endothelial cells and podocytes, promoting lipid deposition in the kidney, and activating local renal inflammatory signaling pathways (such as TLR4/NF‐κB) [32]. Meanwhile, dysfunctional HDL—through the loss of its anti‐inflammatory, antioxidant, and endothelial‐protective functions—renders the kidney more vulnerable to metabolic and inflammatory stress [33]. Therefore, an abnormally increased NHHR may signify a systemic state of lipotoxicity and inflammation, a state that simultaneously erodes both blood vessels and nephrons and accelerates the progression of CKM toward advanced stages. To further deepen the mechanistic discussion of NHHR, it must be situated within the interaction network among the components of metabolic syndrome. Its elevation forms a bidirectional vicious cycle with the core components: visceral obesity and insulin resistance directly raise NHHR by promoting excessive hepatic VLDL synthesis and inhibiting HDL maturation [34]; activation of the renin–angiotensin system associated with hypertension exacerbates oxidative stress, facilitating atherogenic lipoprotein modification and impairing HDL function [35]; and the accumulation of advanced glycation end products (AGEs) induced by hyperglycemia can further worsen lipoprotein quality and function via the AGEs–RAGE axis [14]. Thus, an abnormal rise in NHHR in the early stages of CKM is a sensitive lipid‐level signal of multiple metabolic defects acting in concert, marking the initiation of organ‐damage processes.

Our findings align closely with emerging evidence from multiple populations and disease contexts supporting the predictive value of NHHR for CVD. Yang et al. identified a J‐shaped relationship between NHHR and CVD risk in a U.S. general population, defining an optimal threshold of 2.82, above which CVD risk increased by 21% [36]. Wang et al., using the CHARLS cohort, extended this work by incorporating both baseline and cumulative NHHR, revealing a consistent positive association with CVD incidence [37]. Notably, even among participants with decreasing NHHR trends, those with initially high levels retained an elevated long‐term risk—strongly reinforcing the prognostic value of NHHR. Similarly, in patients with type 2 diabetes, Liu et al. reported that each one‐unit increase in NHHR was associated with a 12% higher risk of adverse cardiovascular events [38]. Together, these studies highlight the universality, dynamic nature, and precision of NHHR as a biomarker of cardiometabolic risk. Consistent with these reports, our ROC curve analysis demonstrated that NHHR had the highest discriminative power (AUC = 0.594) among tested lipid parameters, outperforming both LDL‐C (AUC = 0.534) and non–HDL‐C (AUC = 0.581). Similarly, Yang et al. reported that NHHR exhibited greater sensitivity (66.5%) for CVD prediction than HDL‐C and non–HDL‐C [36].

Notably, in the baseline characteristics of this study, we observed that the proportions of current smokers and current drinkers, as well as cystatin C levels, were lowest in the highest NHHR quartile (Q4). This seemingly paradoxical phenomenon may reflect a “heightened health awareness” effect and more intensive clinical management among high‐risk individuals. Specifically, increasing NHHR levels may prompt both individuals and clinicians to recognize and intervene earlier in cardiometabolic risk, leading to stricter control of smoking and alcohol consumption [20]. In addition, more frequent medical interventions (such as the use of statins and other medications) in this high‐risk group may influence the levels of biomarkers such as cystatin C. Additionally, no significant interaction was found for sex, smoking, drinking, or CKM stage 0–3 (p > 0.05), indicating a consistent NHHR–CVD association across subgroups. Notably, the link was stronger in participants aged <60 years (p‐interaction = 0.0826), suggesting NHHR's particular value for detecting early cardiovascular risk in younger populations. The reason may be that individuals under 60 years old have lower age weighting, with atherosclerosis still in the subclinical stage, making NHHR more likely to “capture” hidden risks; whereas those aged 60 and above present with multiple high‐risk confounders, diluting its incremental predictive power. Collectively, these findings underscore the potential of incorporating NHHR into cardiovascular risk assessment models, with particular utility for early and individualized prevention.

Beyond population‐level associations, the NHHR–CVD relationship may reflect profound molecular and cellular mechanisms. The lipid imbalance represented by elevated NHHR directly contributes to endothelial dysfunction and chronic vascular inflammation [24]. Increased non–HDL‐C components—particularly oxidized LDL (ox‐LDL)—can be recognized by macrophage scavenger receptors such as LOX‐1, CD36 [39], and Toll‐like receptor 4 (TLR4) [40], triggering downstream activation of the NF‐κB signaling pathway and transcription of proinflammatory cytokines [41], including TNF‐α, IL‐6, and MCP‐1.Furthermore, intracellular cholesterol crystal accumulation acts as a “second signal” that activates the NLRP3 inflammasome, leading to caspase‐1 activation and subsequent maturation and secretion of IL‐1β and IL‐18 [42]. These potent inflammatory mediators amplify local immune responses, recruit additional monocytes, and sustain a chronic inflammatory milieu that drives atherosclerotic progression [43]. In the inflammatory and oxidative stress context of CKM, HDL functionality is also impaired [44]. Normally, HDL exerts cardioprotective effects through specific apolipoproteins—especially apoA‐I—and enzymes that mediate cholesterol efflux and endothelial nitric oxide synthase (eNOS) activation [45]. However, under chronic inflammation, HDL particles are modified by serum amyloid A (SAA) and myeloperoxidase (MPO), leading to oxidative nitration of apoA‐I tyrosine residues, thereby compromising cholesterol efflux capacity and endothelial protection [46]. “Acute‐phase” HDL, enriched with SAA, exhibits reduced affinity for macrophage ABCA1 receptors and may even promote endothelial adhesion molecule expression, further exacerbating vascular inflammation [42]. Thus, elevated NHHR may reflect an increase in proinflammatory and dysfunctional HDL particles, representing a collapse of the vascular protective system.

This study has several notable strengths. First, it is the first prospective cohort to systematically evaluate the association between NHHR and incident CVD risk within the framework of CKM syndrome stages 0–3. This investigation expands the clinical applicability of NHHR and emphasizes its value as an early, precision risk marker during the crucial preclinical phase of CKM progression. Second, our analysis leveraged data from the nationally representative CHARLS cohort, characterized by a large sample size, long follow‐up duration, and rigorous methodological quality. The use of comprehensive multivariable adjustment and RCS modeling further enhances the robustness and credibility of the results. Additionally, our extensive subgroup and interaction analyses revealed consistent predictive value across populations, supporting the potential use of NHHR for individualized cardiovascular risk assessment in clinical practice. Nevertheless, several limitations should be acknowledged. First, as an observational cohort study, residual confounding cannot be completely excluded despite extensive covariate adjustment; thus, causality cannot be definitively established. Second, potential measurement bias may exist, as NHHR and certain covariates were based on single baseline measurements, without accounting for longitudinal changes. Furthermore, CVD outcomes were determined through self‐reported physician diagnoses, which, although validated in prior CHARLS‐based studies, may still introduce minor misclassification. Finally, generalizability may be limited: the study population comprised middle‐aged and older Chinese adults, and stringent exclusion criteria might have introduced selection bias. Therefore, external validation using diverse, multiethnic, and multinational prospective cohorts (e.g., UK Biobank, Health and Retirement Study) is warranted to confirm the universality of our findings.

Author Contributions

Yu Huang: contributed to the study's execution, data collection, analysis, interpretation, and initial drafting. Yanwen Gao, Xueming Shao, Ke Wang, Yexiang Ma, and Long Ai: equally contributed to study's interpretation and initial drafting. Jing Yu: presided over the study design, participated in the study process, and manuscript revision.

Funding

This study was supported by the National Natural Science Foundation of China (NSFC 81960086,82160089,82460086), and the Cuiying Scientific and Technological Innovation Program of Lanzhou University Second Hospital (CY2021‐MS‐A13).

Consent

All authors consent to submit the manuscript for publication.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting file 1: jch70221‐sup‐0001‐SuppMat.docx.