The cholesterol, high-density lipoprotein, and glucose (CHG) index as a novel metabolic marker for predicting adverse outcomes in myocardial infarction survivors: insights from two large prospective cohorts

Abstract

Background

Post-myocardial infarction (MI) patients remain at high risk of mortality and recurrent cardiovascular events. Metabolic disorders in patients after MI are closely related to high residual cardiovascular risk. The cholesterol, high-density lipoprotein, and glucose (CHG) index, calculated as Ln {[TC (mg/dL) × FBG (mg/dL)]/[2 × HDL-C (mg/dL)]}, is a recently proposed composite metabolic index. This study aimed to investigate the association between the CHG index and adverse outcomes in MI populations.

Methods

This study included two cohorts: 16,959 individuals with a history of MI from the UK Biobank and 6,253 post-MI patients with coronary artery disease from Fuwai Hospital. The primary endpoints in the UK Biobank cohort were all-cause mortality and cardiovascular mortality. In the Fuwai Hospital cohort, the primary endpoint was major adverse cardiovascular events (MACE, including all-cause mortality, non-fatal MI, and ischemia-mediated revascularization) and hard endpoint (including cardiovascular mortality and non-fatal MI). Cox proportional hazards models, Kaplan–Meier curves, and restricted cubic splines (RCS) were used to evaluate the associations between the CHG index and the endpoints. Time-dependent receiver operating characteristic (ROC) curves were employed to assess the predictive performance.

Results

In the UK Biobank cohort (median follow-up of 13.4 years), after multivariate adjustment, compared to the Q1 of the CHG index, Q4 showed significantly increased risks of all-cause mortality (HR: 1.39, 95% CI: 1.33–1.41) and cardiovascular mortality (HR: 1.42, 95% CI: 1.14–1.74). In the Fuwai Hospital cohort (median follow-up of 3.1 years), the CHG Q4 group also demonstrated a significantly elevated risk of MACE (HR: 1.37, 95% CI: 1.17–1.61) and hard endpoint (HR: 1.87, 95% CI: 1.24–2.81). Kaplan–Meier curves showed significant separation in cumulative event rates across CHG quartiles in both cohorts (log-rank P < 0.05). RCS analyses demonstrated positive linear associations between CHG and all outcomes in both cohorts. Time-dependent ROC curves showed that the CHG index consistently outperformed the TyG index model in predicting adverse outcomes (all FDR-adjusted P < 0.05).

Conclusions

In two large independent cohorts of individuals with prior MI, the CHG index was independently associated with risks of adverse events. While its independent discriminative power is modest, the index serves as a valuable adjunctive tool that enhances risk reclassification, warranting further validation in prospective clinical settings to confirm its utility in secondary prevention.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12933-026-03104-4.

Research insights

What is currently known about this topic?

• •Survivors of myocardial infarction (MI) face high residual cardiovascular risk driven by metabolic disorders.

• •The cholesterol-high-density lipoprotein-glucose (CHG) index is a novel composite metabolic metric, but evidence following MI is limited.

• •The triglyceride-glucose (TyG) index is a well-established marker associated with poor prognosis in post-MI patients.

What is the key research question?

• •Is the novel CHG index associated with adverse outcomes in MI survivors?

• •How does the predictive value of the CHG index compare to that of the TyG index?

What is new?

• •Based on two large prospective cohorts (UK Biobank and Fuwai Hospital), the CHG index is independently related to adverse events in MI survivors.

• •The CHG index shows potential better predictive value of adverse events compared with the TyG index in post-MI populations.

How might this study influence clinical practice?

• •The CHG index may serve as a potential adjunctive tool to enhance post-MI risk stratification.

Introduction

Cardiovascular diseases remain one of the major public health issues causing death and disability worldwide [1]. Patients who have previously experienced a myocardial infarction (MI), even after receiving reperfusion therapy or intervention, have a 12.2% risk of adverse events over three years [2], and the five-year mortality rate reached 37.3% [3]. Therefore, identifying high-risk individuals among patients with a history of myocardial infarction and optimizing prognostic management is a crucial issue that urgently needs to be addressed in the field of cardiovascular medicine.

Metabolic disorders, including insulin resistance (IR), hyperglycemia, dyslipidemia, etc., have been demonstrated to play a critical role in the progression of atherosclerosis, coronary artery disease, and myocardial injury repair [4]. However, single metabolic markers often capture only part of this complex interplay and may not fully reflect the combined burden of lipid and glucose disturbances. Recently, a new composite metabolic indicator—the cholesterol, high-density lipoprotein, and glucose (CHG) index, calculated as Ln {[total cholesterol (TC) (mg/dL) × fasting blood glucose (FBG) (mg/dL)]/[2 × high-density lipoprotein-cholesterol (HDL-C) (mg/dL)]}, has gradually entered the research spotlight [5]. This indicator attempts to more comprehensively reflect the interactive state of lipid and glucose metabolism and holds potential in predicting cardiovascular death or all-cause mortality [6]. However, current research on the CHG index in patients with a history of MI, particularly in terms of prognostic risk assessment, remains very limited.

Therefore, using two large prospective cohorts from the UK Biobank and Fuwai Hospital, the present study evaluated the relationship between the CHG index and post-MI outcomes. It not only addresses the gap in the application of the CHG index among MI survivors but also provides empirical evidence for the potential value of metabolic-lipid composite indicators in the secondary prevention of cardiovascular diseases.

Methods

Study design and participants

This study included a population cohort from the UK Biobank and a clinical cohort from Fuwai Hospital (Fig. 1). The UK Biobank is a large, prospective cohort comprising over 500,000 adults aged 37–73 years recruited from 22 assessment centers across the United Kingdom between 2006 and 2010. At baseline, participants completed detailed questionnaires and interviews on sociodemographic characteristics, lifestyle behaviors, and medical history, and underwent standardized physical and biochemical assessments. All participants provided written informed consent. The study was approved by the North West Multi-Centre Research Ethics Committee, and detailed descriptions of the cohort design have been published previously [7]. Among the 502,367 participants at baseline in UK Biobank, we included participants with a history of MI at recruitment (N = 20,742). Subsequently, we excluded individuals lacking the laboratory data required to calculate the CHG index [TC, HDL-C, and FBG] (n = 3,723), as well as those lost to follow-up (n = 60). The final sample size was 16,959, who were then divided by quartiles of CHG index. (with Q1–Q4 groups consisting of 4,240, 4,240, 4,240, and 4,239 cases, respectively).

Fig. 1.

Flowchart. A UK Biobank data flowchart. B Fuwai Hospital cohort flowchart. CHG cholesterol, high-density lipoprotein, and glucose; FBG fasting blood glucose; HDL-C high-density lipoprotein cholesterol; PCI percutaneous coronary intervention; TC total cholesterol

For Fuwai Hospital cohort, this research involved 30,029 patients with coronary artery disease at Fuwai Hospital, Chinese Academy of Sciences, from January 2017 to December 2018. A total of 8,024 cases with a history of MI were initially included in the screening. After excluding 1,335 cases lacking data on TC, HDL-C, or FBG, as well as 436 cases lost to follow-up, 6,253 patients were ultimately included. All participants were grouped according to CHG quartiles (with Q1–Q4 groups consisting of 1,564, 1,563, 1,563, and 1,563 cases, respectively). The study population consisted exclusively of patients with a history of prior MI who were admitted for elective procedures (such as elective percutaneous coronary intervention) or stable coronary artery assessment, thereby representing a chronic stable phase population. Research adhered to the principles outlined in the Declaration of Helsinki and received approval from the Institutional Review Board at Fuwai Hospital. Prior to the intervention, all participants gave their informed written consent for ongoing follow-up.

Study definitions and outcomes

Based on baseline fasting biochemical data, the CHG index and TyG index were calculated respectively. Fasting venous blood samples were collected on the morning following admission in Fuwai Hospital cohort. Serum TC, HDL-C, triglyceride (TG), and FBG were measured using standard enzymatic methods in certified laboratories. CHG index = Ln {[TC (mg/dL) × FBG (mg/dL)] / [2 × HDL-C (mg/dL)]} [5]. TyG index = Ln [TG (mg/dL) × FBG (mg/dL)/2] [8]. Prior to calculation, FBG, TC, HDL-C, and TG were standardized to mg/dL. Values measured in mmol/L were converted using standard factors: 1 mmol/L = 18 mg/dL for FBG, 1 mmol/L = 38.67 mg/dL for TC and HDL-C, and 1 mmol/L = 88.57 mg/dL for TG. This study chose the TRS2°P score model as the basis for the clinical model [9]. The TRS2°P is a simple risk score incorporating nine clinical characteristics, each is assigned a single point in the total count. These characteristics include age ≥ 75, diabetes mellitus, hypertension, smoking, peripheral artery disease (PAD), previous stroke, previous coronary artery bypass graft surgery (CABG), previous heart failure (HF), and chronic kidney disease (defined by modification of diet in renal disease [MDRD] as < 60 mL/min/1.73m²) [9].

In the UK Biobank, the primary outcomes were all-cause mortality and cardiovascular mortality. Diagnoses were coded according to the International Classification of Disease versions 10 (ICD-10). The specific ICD-10 codes pertaining to MI were I21, I22, I23, I25.2. Participants were monitored until they passed away, or until July 7, 2025. In the Fuwai Hospital cohort, the primary outcome was the occurrence of major adverse cardiovascular events (MACE), including all-cause mortality, non-fatal MI, and ischemia-mediated revascularization. Another major endpoint was the hard endpoint, including cardiovascular mortality and non-fatal MI. The primary endpoints were analyzed as a time-to-first-event outcome. Participants were censored at the occurrence of the first component event. For the analysis of individual components (secondary endpoints), all occurrences of the specific event type were counted, regardless of whether they were the first event. Evaluate the clinical condition of patients at 1, 6, and, 12 months intervals, and subsequently on an annual basis through either outpatient appointments or telephone consultations. All-cause mortality is defined as death from any cause, whether cardiac or non-cardiac. Non-fatal MI was determined based on clinical and laboratory parameters, according to the third universal definition of MI [10]. Ischemia-driven revascularization was defined as any unplanned repeat PCI or CABG of the target lesion or target vessel performed in the presence of recurrent ischemic symptoms, objective evidence of myocardial ischemia, or biomarker/electrocardiogram evidence of ischemia related to the target territory. Planned staged procedures and non–ischemia-driven revascularizations were not counted as events. Any discrepancies in their assessments were addressed by seeking the opinion of a third knowledgeable cardiologist.

Covariables

Covariates were selected based on statistical screening (P < 0.05 in univariate analysis) and clinical relevance (like BMI and smoking) to strictly control for confounding. Meanwhile, we excluded the variables TC, FBG, and HDL-C, which are included in the CHG formula, from the covariates.For the UK Biobank cohort, the following variables were included: age, sex, ethnicity, body mass index (BMI), Townsend deprivation index (TDI), smoking, drinking, metabolic equivalent of task (MET), type 2 diabetes mellitus (T2DM), stroke, low-density lipoprotein cholesterol (LDL-C), hemoglobin A1c (HbA1c), apolipoprotein A (ApoA), lipoprotein (a) [Lp(a)], C-reactive protein (CRP), aspirin, and statins. More detail information was shown in Appendix 1 and Table S1. For Fuwai Hospital cohort, the following variables were included: age, sex, body mass index (BMI), smoking, type 2 diabetes mellitus (T2DM), hypertension, previous HF, previous PVD, previous stroke, high-sensitivity C-reactive protein (hsCRP), LDL-C, Lp(a), HbA1c, dual antiplatelet therapy (DAPT), calcium channel blockers (CCB), β-blockers, and statin. Variance inflation factor (VIF) was presented in Figure S1.

Statistical analysis

Continuous variables were expressed as mean ± standard deviation (SD) if they satisfied normal distribution, otherwise as median [interquartile range (IQR)]. Categorical variables were expressed as percentages (%). Group comparisons were conducted using χ² tests, the Wilcoxon–Mann–Whitney test, or Student’s t-test, as appropriate. For missing data, we imputed the median for variables with a missingness rate < 5%, and treated missing data as a separate category labeled “unknown” for variables with a missingness rate ≥ 5%. Detailed information on the variates with a missingness rate ≥ 5% in the UKB was presented in Table S2.

To examine the relationship between the CHG index and the risk of endpoint events, we established multivariate cox proportional hazards regression models. Univariate Cox regression analysis was also constructed for the covariates. The CHG index was incorporated into the models as continuous and categorical variables (quartiles, Q1–Q4). In UK Biobank population, the quartile ranges for the CHG index were as follows: Q1: ≤ 4.964, Q2: (4.965–5.168], Q3: (5.169–5.402], and Q4: ≥ 5.403. In Fuwai Hospital cohort, the quartile ranges for the CHG index were as follows: Q1: ≤ 4.902, Q2: (4.903–5.026], Q3: (5.027–5.365], and Q4: ≥ 5.366. For the UK Biobank cohort, Model 1 was adjusted for age, sex, and ethnicity (white vs. non-white); Model 2 was further adjusted for BMI, smoking, drinking; Model 3 was further adjusted for MET, TDI, HbA1c, LDL-C, Lp(a), ApoA, T2DM, stroke, and aspirin. For the Fuwai Hospital cohort, Model 1 was unadjusted; Model 2 was adjusted for age, sex, BMI, and smoking; Model 3 was further adjusted for hypertension, previous HF, previous stroke, previous PVD, T2DM, hyperlipidemia, LDL-C, HbA1c, Lp(a), hsCRP, statin, β-blockers, CCB and DAPT. Hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated. The proportional hazards assumption was evaluated using Schoenfeld residuals.

Cumulative incidence curves were plotted using the Kaplan–Meier method and compared via the log-rank test. Multivariable-adjusted restricted cubic spline (RCS) analysis was conducted to evaluate the potential (non-) linear relationship between the CHG index and outcomes.Time-dependent receiver operating characteristic (ROC) analysis was performed to evaluate the predictive accuracy of the CHG index over time. We utilized the Cumulative/Dynamic definition of area under the curve (AUC) proposed by Heagerty et al. [11]. The time-dependent AUCs were calculated at specific time points of 5, 10 and 13 years for UK Biobank cohort, 1 and 3 years for Fuwai Hospital cohort. To account for multiple comparisons in AUC evaluations at various time points, P-values were adjusted using the Benjamini–Hochberg method to control the false discovery rate (FDR) [12]. To evaluate the incremental value of the CHG index, C-statistics, NRI and IDI were calculated. Decision Curve Analysis (DCA) was performed to assess the clinical net benefit and practical utility of the index in risk. Additionally, we performed multivariable Cox regression to compare the independent associations of the CHG index and Non-HDL-C with adverse outcomes. Furthermore, the incremental discriminative power was evaluated by comparing the C-statistics of models integrated with either the CHG index or Non-HDL-C. Spearman correlation was used to assess the relationship between the CHG index and glycemic/lipid biomarkers. To test the robustness of the results, sensitivity analyses were conducted: To ensure the robustness of our findings, several sensitivity analyses were performed: (1) To minimize the potential for reverse causality, we excluded participants who experienced clinical events within the first year of follow-up. (2) To address potential bias from missing data, we excluded covariates with a missingness rate < 5%. (3) Furthermore, covariates with a missingness rate < 5% were handled using Multiple Imputation by Chained Equations (MICE) to ensure data integrity. (4) Finally, to evaluate whether the predictive value of the CHG index was independent of lipid-lowering therapy efficacy, we performed stratified analyses based on LDL-C levels (< 1.8 mmol/L vs. ≥ 1.8 mmol/L), in accordance with clinical guidelines for secondary prevention [13]. In the UK Biobank cohort, subgroup analyses were stratified by age (< 65 years vs. ≥ 65 years), sex (female vs. male), BMI (< 25 vs. ≥ 25 kg/m²) [14], ethnicity (White vs. non-White), TDI (< − 1.56 vs. ≥ 1.56), T2DM (no vs. yes), smoking (no vs. yes), and drinking (no vs. yes). In the Fuwai Hospital cohort, subgroup analyses were performed according to age (< 65 years vs. ≥ 65 years), sex (female vs. male), BMI (< 24 vs. ≥ 24 kg/m²) [15, 16], LVEF (< 50% vs. ≥ 50%), T2DM (no vs. yes), hypertension (no vs. yes), and smoking (no vs. yes). Interaction terms were tested by including cross-product terms in the models. All analyses were performed using R software (version 4.5.1). Two-sided P-values < 0.05 were considered statistically significant.

Results

Baseline characteristics

A total of 16,959 participants from the UK Biobank and 6,253 patients from the Fuwai Hospital cohort were included in the study (Table S3 and Table S4). The median age in the UK Biobank cohort was 63 years, with 76.0% of participants being male. In the Fuwai Hospital cohort, the median age was 59.4 years, and males accounted for 85.6% of the participants. In both cohorts, participants were divided into quartiles based on the CHG index. As the CHG quantile increased, subjects progressively exhibited adverse cardiometabolic characteristics. In the UK Biobank cohort, higher CHG quantiles were strongly associated with significantly elevated BMI, TG, FBG, HbA1c, TC, LDL-C, and CRP, alongside a steady decline in HDL-C and ApoA levels (all P-values < 0.001). The prevalence of diabetes rose markedly from 9.4% in the Q1 group to 44.1% in the Q4 group, while the proportions of current smokers and individuals of non-white ethnicity also showed an increasing trend with higher CHG levels. Similarly, in the Fuwai Hospital cohort, higher CHG quartiles were associated with more adverse metabolic and inflammatory profiles. Compared to the Q1 group, patients in the Q4 group exhibited higher levels of TG, FBG, HbA1c, LDL-C, and hsCRP, but lower levels of HDL-C (all P-values < 0.001). The prevalence of ACS and T2DM significantly increased with ascending quartiles. After reincluding the previously excluded prior MI participants in two cohorts, the baseline distribution remained largely unchanged (Table S5 and Table S6).

Relationship between CHG index and adverse outcome in MI patients

In the UK Biobank cohort, over a median follow-up period of 13.4 years, a total of 5,252 deaths were recorded, including 1,141 cardiovascular deaths. Univariable Cox results for baseline variables are provided in Table S7. We assessed multicollinearity with a threshold of VIF < 5 (Figure S1A). As shown in Table 1, after multivariate stepwise adjustment, higher CHG index values remained consistently associated with an increased risk of all-cause and cardiovascular mortality. In the fully adjusted model (Model 3), compared to the Q1 group, Q4 of the CHG index had an adjusted HR of 1.39 (95% CI 1.33–1.41; P < 0.001) for all-cause mortality and an HR of 1.42 (95% CI 1.14–1.74; P = 0.001) for cardiovascular mortality. Kaplan–Meier survival curves were constructed based on quartiles of the CHG index (Fig. 2). Compared to the lower quartile groups, participants in Q4 had significantly higher cumulative incidences of all-cause mortality and cardiovascular mortality (Log rank test P < 0.05) (Fig. 2A and 2B). Furthermore, multivariable-adjusted RCS indicated an approximately linear association between CHG and both all-cause and cardiovascular mortality (P for overall = 0.0271 and < 0.001; P for non-linearity = 0.428 and 0.113) (Fig. 3A and 3B). In Fuwai Hospital cohort, after a median follow-up of 3.1 years, a total of 729 MACEs were recorded. The distribution of event types contributing to the composite outcomes was detailed in Table S8. Univariable Cox results for baseline variables are provided in Table S9. The VIF for all covariates was less than 5 (Figure S1B). As shown in Table 2, the CHG index was independently associated with MACE (adjusted HR for Q4: 1.37, 95% CI: 1.17–1.61, P < 0.001). Regarding individual components, the CHG index was significantly associated with all-cause mortality (adjusted HR: 1.58, 95% CI: 1.06–2.62, P = 0.029), non-fatal MI (adjusted HR: 1.89, 95% CI: 1.12–3.12, P = 0.025), and ischemia-driven revascularization (adjusted HR: 1.31, 95% CI: 1.07–1.89, P = 0.006). Furthermore, the CHG index was significantly related with hard endpoint (HR 1.87, 95% CI: 1.24–2.81, P = 0.003). Moreover, CHG index displayed significant separation for MACE and hard endpoint (Log rank test P < 0.05) (Fig. 2C and 2D). Multivariable-adjusted RCS indicated a linear increase between CHG and both MACE and hard endpoint (P for overall < 0.001 and = 0.00114; P for non-linearity = 0.576 and 0.922) (Fig. 3C and 3D).

Table 1.

Cox regression model of the association between CHG index and endpoint events in patients with myocardial infarction at UK Biobank database

Event type	Quartile	Total (Events)	Model 1		Model 2		Model 3
			HR (95% CI)	P value	HR (95% CI)	P value	HR (95% CI)	P value
All-cause mortality	CHG index
	Q1	4240(1194)	1.00 [Reference]		1.00 [Reference]		1.00 [Reference]
	Q2	4240(1170)	1.00 (0.92–1.09)	0.969	0.97 (0.90–1.06)	0.510	0.97 (0.99–1.05)	0.397
	Q3	4240 (1272)	1.09 (1.01–1.18)	0.024	1.06 (0.98–1.15)	0.176	1.02 (0.93–1.10)	0.485
	Q4	4239(1616)	1.47 (1.37–1.59)	< 0.001	1.43 (1.32–1.54)	< 0.001	1.39 (1.33–1.41)	< 0.001
	Per 1 unit	16,959 (5252)	1.67 (1.58–1.79)	< 0.001	1.63 (1.54–1.80)	< 0.001	1.54 (1.48–1.84)	< 0.001
Cardiovascular mortality	CHG index
	Q1	4240(207)	1.00 [Reference]	–	1.00 [Reference]	–	1.00 [Reference]	–
	Q2	4240 (256)	1.26 (1.05–1.52)	0.013	1.19 (0.99–1.43)	0.068	1.18 (0.99–1.43)	0.068
	Q3	4240 (281)	1.39 (1.16–1.67)	< 0.001	1.27 (1.06–1.53)	0.009	1.23 (1.02–1.48)	0.036
	Q4	4239 (397)	2.08 (1.76–2.46)	< 0.001	1.87 (1.57–2.22)	< 0.001	1.42 (1.14–1.74)	0.001
	Per 1 unit	16,959 (1141)	2.08 (1.81–2.39)	< 0.001	1.99 (1.71–2.31)	< 0.001	1.41 (1.16–1.73)	< 0.001

Model 1: adjusted for sex, age, ethnicity (white vs non-white). Model 2: additionally adjusted for BMI, smoking, drinking. Model 3: additionally adjusted for MET, TDI, HbA1c, LDL-C, Lp(a), ApoA, CRP, T2DM, stroke, aspirin, and statin. ApoA: apolipoprotein A; BMI body mass index; CRP C-reactive protein; CHG index: cholesterol, high-density lipoprote in, and glucose index; CI: confidence interval; HbA1c hemoglobin A1c; HR hazard ratio; LDL-C low-density lipoprotein cholesterol; Lp(a) lipoprotein(a); MET metabolic equivalent of task; TDI townsend deprivation index; T2DM type 2 diabetes mellitus

Fig. 2.

Kaplan–Meier survival curves for cumulative incidence of endpoint events by CHG index quartiles. Cumulative incidence of all-cause mortality stratified by CHG index quartiles in UK Biobank. B Cumulative incidence of cardiovascular mortality stratified by CHG index quartiles in UK Biobank. C Cumulative incidence of MACE stratified by CHG index quartiles in Fuwai Hospital cohort. D. Cumulative incidence of hard endpoint stratified by CHG index quartiles in Fuwai Hospital cohort. CHG cholesterol, high-density lipoprotein, and glucose; MACE major adverse cardiovascular events

Fig. 3.

Restricted cubic spline (RCS) of CHG index for endpoint events. Restricted cubic spline of CHG index for all-cause mortality in UK Biobank. B. Restricted cubic spline of CHG index for cardiovascular mortality in UK Biobank. C. Restricted cubic spline of CHG index for MACE in Fuwai Hospital cohort. D. Restricted cubic spline of CHG index for hard endpoint Fuwai Hospital cohort. UK Biobank Cohort: adjusted variables same as Model 3 in Table 1. Fuwai Hospital Cohort: adjusted variables same as Model 3 in Table 2. CHG cholesterol, high-density lipoprotein, and glucose; MACE major adverse cardiovascular events; RCS restricted cubic spline

Table 2.

Cox regression model of the association between CHG index and endpoint events in patients with MI at Fuwai Hospital cohort

Event type	CHG quartile	Total (Events)	Model 1		Model 2		Model 3
			HR (95% CI)	P value	HR (95% CI)	P value	HR (95% CI)	P value
MACE
	Q1	1564 (156)	1.00 [Reference]	–	1.00 [Reference]	–	1.00 [Reference]	–
	Q2	1563 (162)	1.06 (0.85–1.32)	0.620	1.13 (0.95–1.34)	0.180	1.14 (0.97–1.42)	0.200
	Q3	1563 (190)	1.24 (1.00–1.53)	0.050	1.26 (1.07–1.49)	0.006	1.21 (1.02–1.51)	0.010
	Q4	1563 (221)	1.44 (1.17–1.77)	< 0.001	1.54 (1.31–1.80)	< 0.001	1.37 (1.17–1.61)	< 0.001
	Per 1 unit	6253 (729)	1.36 (1.17–1.59)	< 0.001	1.34 (1.15–1.57)	< 0.001	1.38 (1.14–1.59)	< 0.001
Hard endpoint
	Q1	1564 (36)	1.00 (Reference)		1.00 (Reference)	–	1.00 (Reference)	–
	Q2	1563 (40)	1.13 (0.79–1.63)	0.501	1.24 (0.86–1.78)	0.254	1.26 (0.87–1.82)	0.226
	Q3	1563 (52)	1.49 (1.06–2.10)	0.022	1.65 (1.17–2.33)	0.004	1.65 (1.14–2.38)	0.008
	Q4	1563 (59)	1.59 (1.14–2.23)	0.006	1.88 (1.34–2.64)	< 0.001	1.87 (1.24–2.81)	0.003
	Per 1 unit	6253 (187)	1.57(1.14–2.01)	< 0.001	1.77 (1.56–2.18)	< 0.001	1.84 (1.34–2.22)	< 0.001
All-cause mortality
	Q1	1564 (41)	1.00 [Reference]	–	1.00 [Reference]	–	1.00 [Reference]	–
	Q2	1563 (42)	1.06 (0.69–1.63)	0.788	1.13 (0.73–1.73)	0.590	1.15 (0.89–1.81)	0.434
	Q3	1563 (65)	1.63 (1.10–2.41)	0.015	1.73 (1.17–2.55)	0.006	1.75 (1.23–2.67)	0.008
	Q4	1563 (57)	1.38 (1.03–2.06)	0.028	1.55 (1.04–2.32)	0.032	1.58 (1.06–2.62)	0.029
	Per 1 unit	6253 (205)	1.41 (1.00–1.88)	0.018	1.52 (1.15–2.01)	0.003	1.59 (1.19–2.25)	0.010
Non-fatal MI
	Q1	1564 (20)	1.00 [Reference]	–	1.00 [Reference]	–	1.00 [Reference]	–
	Q2	1563 (23)	1.23 (0.67–2.27)	0.497	1.27 (0.69–2.33)	0.447	1.20 (0.75–2.10)	0.321
	Q3	1563 (22)	1.18 (0.64–2.18)	0.595	1.21 (0.65–2.23)	0.545	1.18 (0.88–1.98)	0.524
	Q4	1563 (42)	2.19 (1.28–3.77)	0.004	2.26 (1.31–3.89)	0.003	1.89 (1.12–3.12)	0.025
	Per 1 unit	6253 (107)	2.06 (1.42–3.03)	< 0.001	2.09 (1.43–3.05)	< 0.001	1.91 (1.20–3.28)	< 0.001
Ischemia-mediated revascularization
	Q1	1564 (115)	1.00 [Reference]	–	1.00 [Reference]	–	1.00 [Reference]	–
	Q2	1563 (120)	1.05 (0.82–1.36)	0.683	1.06 (0.82–1.36)	0.674	1.10 (0.85–1.74)	0.564
	Q3	1563 (126)	1.10 (0.86–1.42)	0.442	1.11 (0.86–1.42)	0.441	1.14 (0.95–1.67)	0.102
	Q4	1563 (158)	1.39 (1.09–1.77)	0.007	1.39 (1.06–1.77)	0.008	1.31 (1.07–1.89)	0.006
	Per 1 unit	6253 (519)	1.28 (1.07–1.54)	0.008	1.28 (1.06–1.54)	0.008	1.23 (1.06–1.51)	0.001

Model 1: adjusted for none (unadjusted). Model 2: adjusted for age, sex, BMI, and smoking. Model 3: additionally adjusted for hypertension, previous HF, previous stroke, previous PVD, T2DM, hyperlipidemia, LDL-C, HbA1c, Lp(a), hsCRP, statin, β-blockers, CCB and DAPT in addition to age, sex, BMI, smoking. BMI body mass index; CCB calcium channel blocker; CI confidence interval; CHG index cholesterol, high-density lipoprotein, and glucose index; DAPT dual antiplatelet therapy; HbA1c hemoglobin A1c; HR hazard ratio; HF heart failure; hsCRP high-sensitivity C-reactive protein; LDL-C: low-density lipoprotein cholesterol; MACE major adverse cardiovascular events; MI myocardial infarction; PVD peripheral vascular disease; T2DM type 2 diabetes mellitus

Incremental predictive value of the CHG index

To rigorously evaluate the independent prognostic contribution of CHG index, multivariable-adjusted time-dependent ROC curves for the CHG index were conducted (Fig. 4). In UK Biobank, the adjusted AUC for the CHG index model was 0.676 (5-year), 0.673 (10-year), and 0.683 (13-year) for all-cause mortality (Fig. 4A). For cardiovascular mortality, the AUC of CHG index was 0.711 (5-year), 0.694 (10-year), and 0.698 (13-year) (Fig. 4B). In Fuwai Hospital cohort, the AUC of CHG index was 0.697 (1-year), and 0.688 (3-year) for MACE (Fig. 4C). For hard endpoint, the AUC was 0.699 (1-year), and 0.724 (3-year) (Fig. 4D). As detailed in Table 3, the CHG index model consistently achieved higher AUC values compared to the TyG index model at all milestones (all FDR-adjusted P < 0.05). In addition, the addition of the CHG index to the TRS2°P model consistently improved risk discrimination across all endpoints, outperforming both the TRS2°P model alone and the model incorporating the TyG index (Figure S2 and Figure S3). In UK Biobank, for 5-year outcomes, the model with the CHG index achieved the highest AUC for cardiovascular mortality (0.737) and all-cause mortality (0.700), surpassing the predictive value of the TyG index (0.716 and 0.669, respectively) (Figure S2A). The AUC results for 10-year (Figure S2B) and 13-year (Figure S2C) showed similar trends. In the Fuwai Hospital cohort, a similar trend was also observed (Figure S3). Crucially, pairwise comparisons of AUCs confirmed that these improvements remained statistically significant after adjustment for multiple testing using the Benjamini & Hochberg method (all FDR-adjusted P < 0.05). To quantitatively assess whether the CHG index adds prognostic information beyond established risk factors, we calculated the C-statistic improvement, NRI, and IDI (Table S10). In the UK Biobank cohort, the addition of the CHG index to the baseline model significantly improved the C-statistic for all-cause mortality (Δ C-statistic: + 0.022, P < 0.001) and cardiovascular mortality (Δ C-statistic: + 0.027, P < 0.001). Consistent improvements were observed in reclassification metrics, with an NRI of 0.26 and 0.38, and an IDI of 0.018 and 0.030, respectively (all P < 0.001). Similarly, in the Fuwai Hospital cohort, adding the CHG index yielded statistically significant improvements for MACE and hard endpoint (Table S10). Additionally, DCA indicated that the addition of the CHG index to the baseline model yielded a higher net clinical benefit across a reasonable range of threshold probabilities compared to the TyG index and the baseline model alone (Figure S4). We further compared the predictive utility of the CHG index against Non-HDL-C. In Cox regression analysis adjusted for confounding factors, Non-HDL-C failed to show significant associations with adverse outcomes in either cohort (P > 0.05). In contrast, the CHG index remained a significant predictor per 1-SD increase (Table S11). Moreover, the CHG index demonstrated superior discriminative ability, with significantly higher C-statistics compared to Non-HDL-C across all endpoints (Table S12). Spearman correlation analysis showed that the CHG index was moderately and positively correlated with TG, LDL-C, and HbA1c in both cohorts (all P < 0.001), with correlation coefficients (r) ranging from 0.352 to 0.576 (Table S13).

Fig. 4.

Time-dependent ROC curves for the predictive performance of the CHG index. A and B. In the UK Biobank, time-dependent ROC curves were used to measured the predictive value of CHG index for all-cause mortality (A) and cardiovascular mortality (B) at 5 years, 10 years, and 13 years. C and D. In the Fuwai Hospital cohort, time-dependent ROC curves were used to measured the predictive value of CHG index for MACE (C) and hard endpoint (D) at 1 years, and 3 years.. UK Biobank Cohort: adjusted variables same as in Model 3 of Table 1.. Fuwai Hospital Cohort: adjusted variables same as in Model 3 of Table 2. AUC area under the curve; CHG cholesterol, high-density lipoprotein, and glucose; MACE major adverse cardiovascular events; ROC receiver operating characteristic

Table 3.

Predictive performance of CHG index and TyG index across different cohorts and time points

Cohort/Outcome	Time point	Index	AUC (95% CI)	Δ AUC	P value
*UK Biobank cohort
All-cause mortality	5-year	TyG index	0.665 (0.658–0.677)	Ref	–
		CHG index	0.676 (0.691–0.709)	+ 0.011	< 0.001
	10-year	TyG index	0.655 (0.649–0.663)	Ref	–
		CHG index	0.673 (0.667–0.682)	+ 0.018	< 0.001
	13-year	TyG index	0.657 (0.649–0.667)	Ref	–
		CHG index	0.683 (0.665–0.692)	+ 0.026	< 0.001
Cardiovascular mortality	5-year	TyG index	0.690 (0.681–0.705)	Ref	–
		CHG index	0.711 (0.703–0.726)	+ 0.021	< 0.001
	10-year	TyG index	0.669 (0.660–0.678)	Ref	–
		CHG index	0.694 (0.681–0.703)	+ 0.025	< 0.001
	13-year	TyG index	0.671 (0.662–0.679)	Ref	–
		CHG index	0.698 (0.689–0.607)	+ 0.027	< 0.001
#Fuwai Hospital
MACE	1-year	TyG index	0.682 (0.662–0.702)	Ref	–
		CHG index	0.697 (0.677–0.717)	+ 0.015	0.028
	3-year	TyG index	0.675 (0.655–0.695)	Ref	–
		CHG index	0.688 (0.668–0.708)	+ 0.013	0.035
Hard endpoint	1-year	TyG index	0.670 (0.650–0.690)	Ref	–
		CHG index	0.699 (0.679–0.719)	+ 0.029	0.005
	3-year	TyG index	0.700 (0.680–0.720)	Ref	–
		CHG index	0.724 (0.704–0.744)	+ 0.024	0.002

^*Adjusted for age, sex, ethnicity (white vs non-white), BMI, drinking, smoking, MET, TDI, HbA1c, Lp(a), LDL-C, ApoA, CRP, T2DM, aspirin, and statin

^#Adjusted for age, sex, BMI, hypertension, previous HF, previous stroke, previous PVD, T2DM, hyperlipidemia, HbA1c, Lp(a), LDL-C, hsCRP, statin, β-blockers, CCB and DAPT

ApoA apolipoprotein A; BMI body mass index; CCB calcium channel blocker; CRP C-reactive protein; CHG index cholesterol, high-density lipoprote in, and glucose index; CI confidence interval; DAPT dual antiplatelet therapy; HbA1c hemoglobin A1c; HR hazard ratio; HF heart failure; hsCRP: high-sensitivity C-reactive protein; LDL-C: low-density lipoprotein cholesterol; Lp(a): lipoprotein(a); MACE: major adverse cardiovascular events; MI: myocardial infarction; MET metabolic equivalent of task; PVD peripheral vascular disease; TDI townsend deprivation index; T2DM type 2 diabetes mellitus; TyG: triglyceride–glucose

Sensitivity analyses and subgroup analyses

We conducted several prespecified sensitivity analyses: (1) excluding participants with a missingness rate < 5% covariates (Tables S14 and S15); (2) using MICE to handle missing data < 5% (Tables S16 and S17); (3) excluding events within 1 year of baseline (Tables S18 and S19); and (4) stratifying patients by LDL-C levels (< 1.8 mmol/L vs. ≥ 1.8 mmol/L) (Tables S20). The results were similar in trend to the main results. As shown in the Table S20, the CHG index remained a significant and independent predictor of adverse outcomes in both the LDL-C < 1.8 mmol/L and LDL-C ≥ 1.8 mmol/L subgroups across both cohorts. The results of subgroup analyses are summarized in Table S21 and Table S22. In the UK Biobank cohort (Table S21), the association between higher CHG and all-cause mortality as well as cardiovascular mortality showed consistent directionality across all subgroups (age, sex, BMI, ethnicity, TDI, T2DM, smoking, and drinking). In the Fuwai Hospital cohort (Table S22), the association between CHG and MACE remained stable across all subgroups (age, BMI, LVEF, T2DM, hypertension, and smoking). A sex-based interaction was observed (interaction P-value = 0.034): the association was stronger in males (HR = 1.18, 95% CI 1.05–1.32) than in females (HR = 0.94, 95% CI 0.73–1.20).

Discussion

In this study, we evaluated the association between CHG index and adverse events in patients with prior MI based on two large-scale, prospectively designed cohorts (the UK Biobank and Fuwai Hospital cohorts). The main findings are as follows: (1) Among patients with prior MI, CHG index shows an independent linear positive correlation with the risk of adverse events. (2) The CHG index has moderate predictive capability as an independent indicator and outperforms TyG. (3) The CHG index can enhance the predictive performance of the existing clinical model TRS2°P score.

Patients with a history of MI are classified as a very high-risk group, and their long-term prognosis is influenced by multiple interrelated factors. Metabolic disorders—including abnormalities in glucose and lipid metabolism and IR—have been demonstrated to be significant determinants of adverse outcomes [17–19]. Abnormal glucose metabolism significantly increases the risk of death in MI patients: The short-term and long-term mortality risk for MI patients with diabetes is more than twice that of non-diabetic patients, indicating that hyperglycemia and glucose metabolism disorders are major risk factors following MI [20]. Dyslipidemia also have a profound impact on the prognosis [21, 22]. Elevated LDL-C and reduced HDL-C are traditional risk factors for atherosclerosis and coronary heart disease. Although treatments such as statins have significantly reduced the recurrence risk in MI patients, residual risk remains [23]. IR is a potential trigger for cardiovascular diseases, often accompanied by manifestations such as hyperglycemia, hyperinsulinemia, elevated blood pressure, and dyslipidemia [24–26]. These metabolic abnormalities, when combined, can promote vascular endothelial dysfunction, inflammatory responses, and thrombosis, ultimately leading to poorer cardiovascular outcomes [27–29]. Therefore, in individuals with a history of MI, identifying indicators that comprehensively reflect glucose and lipid metabolism is of significant importance for clinical prognosis assessment.

The CHG index is a novel composite metabolic indicator. Mansoori et al. [5] first proposed the CHG index for the identification of T2DM and showed that it is highly correlated with IR and abnormal glucose metabolism, with diagnostic performance superior to that of the traditional TyG index, suggesting that incorporating cholesterol-related indicators can enhance the assessment of metabolic disorders. Subsequently, CHG has been applied to cardiovascular risk prediction. Mo et al. [30], using data from the CHARLS cohort of older adults in China, compared the predictive ability of CHG and TyG for cardiovascular events and found that an increase in CHG was linearly and positively associated with cardiovascular risk, with predictive ability comparable to that of TyG. In critically ill patients with ischemic stroke, an elevated CHG index was significantly associated with 28-day in-hospital mortality, with prognostic performance similar to that of TyG [31]. In patients with calcific aortic stenosis, the ARISTOTLE study further demonstrated that an elevated CHG index independently predicts cardiovascular death and all-cause mortality [32]. Overall, although existing evidence is still limited, it consistently supports a close association between the CHG index, metabolic abnormalities, and cardiovascular risk. Our findings extend these observations to post-MI patients, showing that higher CHG is significantly associated with adverse events, suggesting a potential role of CHG in the secondary prevention of coronary heart disease. A critical concern in post-MI management is the extent to which metabolic markers identify residual risk independent of standard lipid-lowering therapy. Our sensitivity analysis stratified by LDL-C levels offers compelling evidence in this regard. Notably, the CHG index remained a robust and independent predictor of cardiovascular mortality and hard endpoints even among patients who achieved the stringent secondary prevention target of LDL-C < 1.8 mmol/L. This suggests that the predictive value of the CHG index is not merely a reflection of poor statin compliance or sub-optimal dosing, which might otherwise manifest as elevated LDL-C. Instead, the CHG index likely captures a unique dimension of ‘residual cardiometabolic risk’ by integrating the synergistic effects of glucose dysregulation and atherogenic non-HDL-C particles. These findings underscore the clinical utility of the CHG index in identifying high-risk individuals who may require intensified metabolic intervention even when their traditional LDL-C targets appear well-controlled. Furthermore, it is essential to compare the CHG index with established lipid markers such as Non-HDL-C, which is currently emphasized in secondary prevention guidelines. Our results indicated that the CHG index is not merely a mathematical variation of Non-HDL-C but a more potent predictor of cardiovascular risk. In our head-to-head comparison, the CHG index consistently demonstrated superior discriminative ability, with significantly higher C-statistics and independent associations compared to Non-HDL-C alone. This superiority likely stems from the index’s ability to integrate both the cumulative atherogenic burden of lipid particles and the deleterious effects of hyperglycemia. In patients with a prior MI, the interplay between dyslipidemia and glucose intolerance often leads to a heightened state of chronic inflammation and vascular endothelial dysfunction. By capturing this synergistic relationship, the CHG index provides a more comprehensive metabolic profile, justifying its slightly increased complexity in exchange for significantly enhanced clinical predictive utility in high-risk populations.

It is noteworthy that although the endpoint settings in the two cohorts of this study differ, this discrepancy precisely enhances the generalizability of the research findings. The UK Biobank, as a large population-based prospective study with a long follow-up period (median 13.4 years), primarily focuses on all-cause mortality and cardiovascular mortality as its main endpoints. This provides excellent evidence for evaluating the long-term life expectancy predictive value of the CHG index. In contrast, the Fuwai Hospital cohort (median follow-up 3.1 years) emphasizes short-term clinical risks after MI. We observed that the CHG index was significantly associated not only with the composite MACE but also with each of its individual components, including all-cause mortality, non-fatal MI, and ischemia-mediated revascularization. This consistency indicates that the CHG index reflects a widespread increase in systemic cardiovascular risk, rather than being disproportionately driven by a single, low-weight clinical event. Notably, the association remained consistent when a more stringent “hard endpoint” (defined as cardiovascular mortality and non-fatal MI) was utilized. This finding is of particular clinical importance, as hard endpoints are less susceptible to elective clinical decision-making and more directly reflect the severe pathophysiological consequences of metabolic dysregulation after MI. In addition, it is important to note that we utilized the TRS2°P score as the baseline model, as it is specifically validated for long-term risk stratification in patients with prior MI [9] (unlike GRACE [33] or TIMI [34], which are intended for acute coronary syndromes [35–37]). The effectiveness of TRS2°P in secondary prevention prediction for MI patients has been validated by multiple studies [38–40]. The addition of the CHG index to the TRS2°P-based model yielded a significant improvement in the C-statistic and resulted in a substantial NRI. This indicated that the CHG index shows potential as a complementary metabolic marker for risk assessment and warrants further validation in prospective clinical settings.

The TyG index is a simple and widely used indicator of IR, calculated from TG and FBG levels. [41] Previous studies have shown that the TyG index is an important metabolic prognostic predictor in patients with MI. [42–44] Higher TyG has been associated with both short-term in-hospital outcomes and long-term outcomes. However, some studies have suggested that its predictive power is relatively limited in MI patients without diabetes. [45] Against this background, we found that, in predicting prognosis among MI survivors, the CHG index performs slightly better than the TyG index. This finding differs from the “comparable efficacy” reported by Mo et al. [30] and suggests that in very high-risk subgroups such as MI survivors, a metabolic index incorporating cholesterol may better capture subtle risk differences than a triglyceride-based index alone.

The potential mechanisms underlying the predictive ability of the CHG index may be related to both the pathophysiological significance of its components and their synergistic effects. Our Spearman correlation analysis provided a data-driven foundation for this link, showing that the CHG index is moderately and positively correlated with TG, LDL-C, and HbA1c across both cohorts (r: 0.352–0.576; all P < 0.001). These findings suggest that while the CHG index is biologically anchored in established metabolic pathways, it integrates multidimensional risk information that individual biomarkers may not fully capture. First, we hypothesize that the glucose component reflects disturbances in glucose homeostasis and IR. Hyperglycemia exacerbates cardiovascular damage after MI through oxidative stress, endothelial dysfunction, inflammation, and enhanced platelet aggregation, adversely affecting myocardial remodeling, plaque stability, and endothelial repair [46–49]. Second, the TC/HDL-C components capture core features of atherogenic lipid abnormalities: high TC, reflecting the burden of atherogenic lipoproteins (including LDL and cholesterol remnants), provides the substrate for plaque formation and progression, whereas low HDL-C indicates impaired reverse cholesterol transport and reduced anti-inflammatory, antioxidant, and antithrombotic functions [50–52]. Moreover, the CHG index reflects the interaction of these two pathological states. IR and hyperglycemia exacerbate atherogenic dyslipidemia (such as increased hepatic VLDL production and reduced HDL-C levels), while lipotoxicity in turn aggravates IR, forming a vicious cycle [5, 49, 53]. Although the TyG index includes TG, the pathogenicity of TG is largely mediated through its remnants and its close association with low HDL-C [54]. By directly incorporating TC and HDL-C, the CHG index may more comprehensively reflect lipoprotein-related risk. Thus, CHG integrates the two core metabolic axes driving residual risk after MI—glucose toxicity and lipotoxicity—which may explain why it retains prognostic value and slightly outperforms TyG. Additionally, since the CHG index integrates TC and HDL-C, it implicitly captures the components of Non-HDL-C, which is closely related to the accumulation of TG-rich lipoprotein remnants (TRL remnants) [55, 56]. In the context of IR and hyperglycemia commonly seen in patients after MI, impaired lipolysis of very low-density lipoprotein (VLDL) leads to the accumulation of these highly atherogenic remnants, which are more prone to penetrate the vascular intima and induce inflammation than LDL [57, 58]. This explains why the CHG index has stronger prognostic predictive power compared to single Non-HDL-C .

This study has several limitations. First, as an observational cohort analysis, despite extensive adjustments, residual confounding factors cannot be entirely ruled out, and causality cannot be established. Whether a higher CHG index directly exerts mechanistic effects or merely reflects underlying metabolic risks remains to be elucidated. Second, only baseline CHG values were available, and dynamic changes in metabolic indicators during follow-up could not be assessed. Future research should incorporate repeated measurements or trajectory analysis. Third, the UK Biobank primarily enrolled relatively healthy European populations, whereas the Fuwai Hospital cohort consisted of Chinese patients with confirmed coronary heart disease. Although the results were consistent across both cohorts, differences in ethnicity, disease spectrum, and baseline risks may limit the generalizability of the findings, necessitating exploring in more diverse populations. Fourth, although the CHG index demonstrates significant incremental predictive value, its independent discriminative ability in clinical settings remains limited. Fifth, the follow-up durations differed significantly between the two cohorts (approximately 3 years versus 10 years), which may affect the comparability of predictive performance—shorter follow-up could underestimate risk associations, while longer follow-up is susceptible to influences from evolving clinical management practices. The impact of follow-up duration on the prognostic value of CHG warrants further investigation. Finally, gender-specific differences in the Fuwai cohort was observed. Unlike the population-based UK Biobank, the Fuwai Hospital cohort is a clinically-based study that continuously recruits patients upon their admission for elective surgery or stable coronary artery evaluation and treatment. In such a real-world clinical setting, gender imbalance is often unavoidable [59, 60]. Future large-scale, gender-balanced clinical studies are needed to further clarify the application value of the CHG index for female patients after MI in the Chinese clinical context.

Conclusion

In this study based on two independent MI cohorts, the CHG index was demonstrated to be an independent predictor of adverse outcomes. The CHG index may serve as a potential adjunctive tool to enhance post-MI risk stratification, and warrants further investigation as a potential adjunctive marker.

Abbreviations

ACS

Acute coronary syndrome

ApoA

Apolipoprotein A

AUC

Area under the curve

BMI

Body mass index

CABG

Coronary artery bypass grafting

CCB

Calcium channel blockers

CCS

Chronic coronary syndrome

CHG

Cholesterol–high-density lipoprotein–glucose

CI

Confidence intervals

CRP

C-reactive protein

CTO

Chronic total occlusion

CVD

Cardiovascular diseases

DAPT

Dual antiplatelet therapy

DBP

Diastolic blood pressure

DM

Diabetes mellitus

FBG

Fasting blood glucose

HbA1c

Hemoglobin A1c

HDL-C

High-density lipoprotein-cholesterol

HF

Heart failure

HR

Hazard ratios

hsCRP

High-sensitivity C-reactive protein

ICD-10

International classification of disease versions 10

IQR

Interquartile range

IR

Insulin resistance

LAD

Left anterior descending

LCX

Left circumflex

LDL-C

Low-density lipoprotein cholesterol

LM

Left main

Lp(a)

Lipoprotein (a)

LVEF

Left ventricular ejection fraction

MACE

Major adverse cardiovascular events

MET

Metabolic equivalent of task

MI

Myocardial infarction

PCI

Percutaneous coronary intervention

PVD

Peripheral vascular disease

RCA

Right coronary artery

ROC

Receiver operating characteristic

SBP

Systolic blood pressure

SD

Standard deviation

T2DM

Type 2 diabetes mellitus

TC

Total cholesterol

TDI

Townsend deprivation index

TG

Triglycerides

TyG

Triglyceride-glucose

Author contributions

KD, YS, and XC designed the research; YS and XC performed the statistical analysis; ZC, XB, JH, BL, ZZ, CW, ZL, CZ and RF contributed to data collection and interpretation; XC and YS wrote the paper; YS, XC, RF and KD revised the manuscript. All authors read and approved the final manuscript.

Funding

This work was supported by Noncommunicable Chronic Diseases-National Science and Technology Major Project (2025ZD0548200) .

Data availability

Researchers interested in accessing the data used in this study can apply for access to the UK Biobank by visiting their website (https://www.ukbiobank.ac.uk/enable-your-research/apply-for-access) and submitting an application that includes a research protocol summary and requested data fields. Upon approval by the UK Biobank management team and payment of applicable fees, researchers will be granted access to the dataset.

Declarations

Ethics approval and consent to participate

UK Biobank has been approved by the National Research Ethics Committee of the National Health Service (NHS). All participants signed informed consent.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.