Impact of Glucose Metabolism Status on the Association between Apolipoprotein A-I and Ischemic Risk in Patients with Coronary Artery Disease: A Large-Sample Cohort Study

Study design and population

In 2013, a total of 10,724 consecutive CAD patients who underwent PCI at Fu Wai Hospital (National Center for Cardiovascular Diseases, Beijing, China) were enrolled. Patients with missing key baseline laboratory data—specifically glycated hemoglobin (HbA1c), fasting plasma glucose (FPG), or ApoA-I—were excluded from the analysis. Ultimately, 10,232 patients were included in the main analysis (Fig. 1).

Fig. 1.

Patient enrollment flowchart in the cohort. PCI, percutaneous coronary intervention; HbA1c, glycated hemoglobin; FPG, fasting plasma glucose; ApoA-I, apolipoprotein A-I.

Before PCI, all participants received aspirin and a P2Y12 inhibitor. A loading dose of 300 mg of aspirin and either 300 mg of clopidogrel or 180 mg of ticagrelor was administered to those who had not previously received antiplatelet therapy. Following PCI, patients were prescribed 100 mg of aspirin daily, along with either clopidogrel (75 mg, once daily) or ticagrelor (90 mg, twice daily) for at least 1 year.

All procedures involving human participants were conducted in accordance with institutional and international ethical standards, and all individuals provided informed consent. The study was approved by the Ethics Committee of Fu Wai Hospital (Approval Number: 2013-449), in compliance with the Declaration of Helsinki.

Data measurements and definitions

Blood samples were collected the morning after an overnight fast, within 24 hours of hospital admission. Apolipoprotein levels, including ApoA-I and apolipoprotein B, were measured using the immunoturbidimetric method (apolipoprotein A1 FS and apolipoprotein B FS multi-purpose kits; DiaSys Diagnostic Systems GmbH, Holzheim, Germany). Fasting glucose was measured with commercially available test kits (Biosino Bio-Technology and Science Incorporation, Beijing, China) using the glucose oxidase technique. HbA1c levels were determined by high-performance liquid chromatography. All laboratory assessments were performed at the core laboratory of Fu Wai Hospital using standard biochemical techniques. Baseline characteristics—including demographics, medical history, admission diagnoses, laboratory results, procedural details, and discharge medications—were comprehensively recorded.

Glucose metabolism status was defined according to the latest American Diabetes Association criteria [17]. Specifically, (1) DM was diagnosed if any of the following were present: HbA1c ≥6.5% (48 mmol/mol), FPG ≥7.0 mmol/L (126 mg/dL), 2-hour plasma glucose during oral glucose tolerance test (OGTT) ≥11.1 mmol/L (200 mg/dL), a previous diagnosis of diabetes, or current or prior use of oral hypoglycemic agents or insulin; (2) Pre-DM was defined as the absence of self-reported diabetes, with HbA1c between 5.7% and 6.4% (39 to 47 mmol/mol), FPG between 5.6 and 6.9 mmol/L (100 to 125 mg/dL), or 2-hour plasma glucose during OGTT between 7.8 and 11.0 mmol/L (140 to 199 mg/dL); (3) NGR was defined as not meeting the criteria for DM or pre-DM.

Clinical endpoints and follow-up

The primary endpoint was the occurrence of major adverse cardiac and cerebrovascular events (MACCE), which included all-cause death, myocardial infarction (MI), stroke, and coronary revascularization. Each component of the MACCE composite endpoint was also analyzed separately. MI was diagnosed according to the third universal definition of MI [18]. Revascularization was defined as unplanned repeat revascularization for ischemic symptoms or events, performed by either PCI or coronary artery bypass grafting. Stroke was diagnosed according to the most recent relevant guidelines [19]. Patients were followed up at 30 days, 6 months, 1 year, 2 years, and 5 years after discharge, achieving a follow-up rate of 91.5% at 5 years. Follow-up was conducted either via hospital visits or telephone interviews, with the occurrence of clinical events systematically documented. Time to event was defined as the duration from PCI to the event of interest, all-cause death, or the date of last follow-up, whichever occurred first. All adverse events were meticulously recorded, verified, and adjudicated by a panel of independent clinical cardiologists.

Statistical analysis

Normally distributed continuous variables were presented as mean±standard deviation, while non-normally distributed variables were reported as median with interquartile range. Categorical variables were expressed as counts with percentages. Comparisons of continuous variables between groups were assessed using Student’s t test, Mann-Whitney test, or analysis of variance (ANOVA), as appropriate; categorical variables were compared using the chi-square test or Fisher’s exact test. Based on diagnostic criteria for glucose metabolism status, patients were categorized into DM, pre-DM, and NGR groups. ApoA-I was analyzed both as a continuous variable and as a categorical variable, according to quintiles defined separately for the overall cohort as well as within the DM, pre-DM, and NGR groups (Supplemental Table S1).

For the overall cohort, the continuous association between ApoA-I and MACCE risk was evaluated using restricted cubic spline (RCS) analysis, with the median ApoA-I level in each group serving as the reference. Kaplan-Meier survival curves and the log-rank test were utilized to analyze cumulative incidences of MACCE across different ApoA-I quintiles. Univariate (model 1) and multivariate (model 2) Cox proportional hazards regression models were employed to further explore associations between ApoA-I quintiles and clinical outcomes. Multivariate Cox regression model (model 2) was adjusted for significant variables in the univariate Cox model for MACCE (Supplemental Table S2). Hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated. P values for interaction were computed to assess whether the association between ApoA-I and clinical outcomes differed according to glucose metabolism status. To further evaluate these interactions, all analyses described above were conducted both in the overall cohort and within each glucose metabolism group separately.

For subgroup analyses, the overall cohort was stratified by sex (men and women), age (subgroup analyses, based on the World Report on Ageing and Health [20]), and BMI (on Ageing and Hea according to Chinese criteria for overweight and obesity [21]). Sensitivity analyses were performed to assess the effects of demographic and lifestyle factors (model 3), comorbid diseases, clinical diagnosis and revascularization information (model 4), key lipid variables (model 5), and cardiovascular medications as well as insulin therapy (model 6; insulin was adjusted only in the DM group) (Supplemental Table S3). Additionally, participants were classified into six mutually exclusive groups based on ApoA-I dichotomy (using the median level as the cutoff) and glucose metabolism status to assess their combined effect on MACCE: ApoA-I ≥ median or < median within NGR; ApoA-I ≥ median or < median within pre-DM; ApoA-I ≥ median or < median within DM. A two-sided P value <0.05 was considered statistically significant for all analyses. All statistical analyses were performed using SPSS software version 26.0 (IBM Corp., Armonk, NY, USA) and R software version 4.2.3 (R Foundation for Statistical Computing, Vienna, Austria).

Availability of data and materials

Due to ethical restrictions related to the consent given by subjects at the time of study commencement, our datasets are available from the corresponding author upon reasonable request after permission of the Institutional Review Board of State Key Laboratory of Cardiovascular Disease, Fu Wai Hospital, National Center for Cardiovascular Diseases.

Baseline characteristics

A total of 10,232 consecutive patients were included in the main analysis after excluding those who met the exclusion criteria (Fig. 1). The mean age was 58.37±10.29 years, and 7,895 (77.16%) were male. Of these, 4,602 (44.98%) had DM (all type 2), 4,498 (43.96%) had pre-DM, and 1,132 (11.06%) had NGR. The median ApoA-I level in the overall population was 1.31 g/L. Over a 5-year follow-up, 2,139 (20.90%) patients experienced MACCE.

Patients who experienced MACCE tended to be older and had a higher prevalence of hypertension and chronic obstructive pulmonary disease. They also had higher rates of prior MI, previous stroke, prior PCI, and previous coronary artery bypass grafting. Additionally, they had elevated FPG, HbA1c, and lipoprotein (a) levels, and lower estimated glomerular filtration rate and left ventricular ejection fraction (Table 1).

Table 1.

Baseline Characteristics of Patients with and without MACCE

Baseline characteristics stratified by glucose metabolism status are shown in Supplemental Table S4, and those of patients divided into quintiles of ApoA-I levels are shown in Supplemental Table S5.

Association between ApoA-I levels and clinical outcomes in the overall cohort

For MACCE, RCS analysis demonstrated a significant correlation between continuous ApoA-I levels and MACCE (P for overall=0.033), with lower levels of ApoA-I significantly increasing the risk of MACCE (Fig. 2). When ApoA-I was analyzed as a categorical variable (quintiles), patients in the lowest quintile (Q1) had the highest incidence of MACCE (22.7%), whereas those with ApoA-I levels between 1.37 and 1.52 g/L (Q4) had the lowest incidence (19.8%). After adjustment for risk factors, multivariate Cox regression (model 2) found a 1.207-fold increased risk of MACCE in patients with the lowest ApoA-I quintile (HR, 1.207; 95% CI, 1.054 to 1.380; P=0.006) (Fig. 3). The P value for interaction for continuous ApoA-I levels and glucose metabolism status was 0.041; and that for the interaction for ApoA-I quintiles and glucose metabolism status was 0.047, both suggesting that glucose metabolism status modified the relationship between ApoA-I and MACCE.

Fig. 2.

Restricted cubic spline analysis of the relationship between continuous apolipoprotein A-I (ApoA-I) levels and major adverse cardiac and cerebrovascular event (MACCE). Adjusted for age, hypertension, chronic obstructive pulmonary disease, previous myocardial infarction, previous stroke, prior percutaneous coronary intervention, prior coronary artery bypass grafting, lipoprotein (a), estimated glomerular filtration rate, left ventricular ejection fraction, glucose, and glycated hemoglobin A1c. Curves represent the natural log of hazard ratios (HRs) of ApoA-I levels along a continuous spectrum, with the median value in each group being the reference. (A) Overall cohort, (B) diabetes mellitus (DM) group, (C) pre-DM group, and (D) normal glucose regulation (NGR) group. Shaded areas represent the 95% confidence interval (CI). ^aP values indicating statistical significance.

Fig. 3.

Cox regression for apolipoprotein A-I (ApoA-I) quintile and major adverse cardiac and cerebrovascular event (MACCE). Model 2 was adjusted for age, hypertension, chronic obstructive pulmonary disease, previous myocardial infarction, previous stroke, prior percutaneous coronary intervention, prior coronary artery bypass grafting, lipoprotein (a), estimated glomerular filtration rate, left ventricular ejection fraction, glucose, and glycated hemoglobin A1c. P value for interaction: glucose metabolism statuses (categorical)×ApoA-I (quintiles)=0.047. HR, hazard ratio; CI, confidence interval; NA, not applicable; DM, diabetes mellitus; NGR, normal glucose regulation. ^aP values indicating statistical significance.

For the individual components of MACCE, the lowest ApoA-I quintile was consistently associated with the highest incidence rates of all-cause death, MI, and revascularization (Supplemental Figs. S1-S4). Notably, multivariate Cox regression (model 2) showed that the lowest ApoA-I quintile was independently associated with a higher risk of all-cause death (HR, 1.581; 95% CI, 1.123 to 2.226; P=0.009).

Association between ApoA-I and clinical outcomes for different glucose metabolism statuses

RCS analysis showed that, for MACCE, only in patients with DM did ApoA-I display a non-linear association, with low levels significantly increasing MACCE risk (P for non-linearity=0.044); no association was seen in pre-DM or NGR groups (both P>0.05) (Fig. 2). When ApoA-I was categorized into quintiles, Kaplan-Meier analysis indicated a significant difference in cumulative MACCE incidence across quintiles only in the DM group (log-rank P=0.040), but not in the pre-DM or NGR group (Supplemental Fig. S5). Subsequent multivariate Cox regression (model 2) further confirmed that ApoA-I quintiles were not associated with MACCE in the pre-DM or NGR groups. However, in the DM group, the lowest ApoA-I quintile (Q1: <1.14 g/L) was associated with a 1.327-fold increased risk of MACCE compared to those at lowest risk (HR, 1.327; 95% CI, 1.097 to 1.604; P=0.003) (Fig. 3).

For MACCE components in the DM group, patients in the lowest ApoA-I quintile had the highest incidence of all-cause death, MI, and revascularization (Supplemental Figs. S1-S4). Further multivariate Cox regression (model 2) showed that lower ApoA-I was independently associated with increased risk of MI (Q1: HR, 1.571; 95% CI, 1.044 to 2.364; P=0.030) and stroke (Q2: HR, 1.693; 95% CI, 1.049 to 2.730; P=0.031). In the pre-DM group, lower ApoA-I (Q2) was linked to a higher risk of stroke (HR, 1.856; 95% CI, 1.057 to 3.258; P=0.031). No significant associations were observed between ApoA-I quintiles and individual MACCE components in the NGR group.

Subgroup analysis

Subgroup analyses stratified by sex, age, and BMI showed that the association between the lowest ApoA-I quintile and higher MACCE risk, specifically in the DM group, remained consistent across all subgroups (Fig. 4).

Fig. 4.

Association between lowest apolipoprotein A-I (ApoA-I) quintile and major adverse cardiac and cerebrovascular event (MACCE) at different glucose metabolism statuses in subpopulations. Adjusted for age, hypertension, chronic obstructive pulmonary disease, previous myocardial infarction, previous stroke, prior percutaneous coronary intervention, prior coronary artery bypass grafting, lipoprotein (a), estimated glomerular filtration rate, left ventricular ejection fraction, glucose, and glycated hemoglobin A1c. Hazard ratios (HRs) and confidence intervals (CIs) show the association between the lowest quintile of ApoA-I levels and MACCE. DM, diabetes mellitus; NGR, normal glucose regulation; BMI, body mass index. ^aP values indicating statistical significance.

Sensitivity analyses

After additional adjustment for demographic and lifestyle factors (model 3), comorbidities, clinical diagnoses, revascularization data (model 4), key lipid variables (model 5), and cardiovascular medication and insulin therapy (model 6, with insulin only adjusted in the DM group) in multivariate Cox regression, the association between the lowest ApoA-I quintile and increased MACCE risk remained robust in both the overall and DM groups (Supplemental Table S6). Results for MACCE components are presented in Supplemental Tables S7-S10.

Combined effect of ApoA-I and diabetes for MACCE

A combined analysis was conducted to evaluate the joint association of ApoA-I and diabetes status with MACCE. Stratifying patients by ApoA-I median and glucose metabolism status, Kaplan-Meier curves demonstrated that patients with both low ApoA-I (< median) and DM status had the highest cumulative incidence of MACCE (Supplemental Fig. S6). Multivariate Cox regression showed that, compared with those having higher ApoA-I and NGR status, patients with both low ApoA-I (< median) and DM status had the highest risk of MACCE (HR, 1.288; 95% CI, 1.003 to 1.653) (Table 2).

Table 2.

Cox Regression Analysis for MACCE Risk Stratified by Dichotomized ApoA-I Level and Glucose Metabolism Status

Lower ApoA-I and adverse outcomes in CAD patients

ApoA-I plays an indispensable role in lipid metabolism. Although numerous studies have examined the association between ApoA-I and adverse outcomes across various populations, their conclusions remain inconsistent. Furthermore, research specifically addressing ApoA-I–related ischemic risk in PCI patients is limited.

In populations without known CAD, a study demonstrated a linear, negative correlation between ApoA-I and fatal MI [22], whereas a more recent study by Faaborg-Andersen et al. [23] found a more complex, U-shaped association between ApoA-I levels and mortality risk, suggesting that both very low and very high ApoA-I levels may increase mortality risk. In CAD patients, a small-sample study by Garfagnini et al. [24] involving 140 individuals with acute MI confirmed an inverse association between ApoA-I and CAD severity. In PCI-treated CAD patients, Nishiyama et al. [25] studied 3,835 PCI patients and found that lower ApoA-I levels were associated with an increased risk of cancer death.

Compared to these previous studies, our work stands out for its large sample size of PCI patients, providing more robust evidence for the relationship between reduced ApoA-I levels and poor prognosis in this high-risk CAD population. Notably, our study used MACCE as the clinical endpoint, offering comprehensive consideration of a range of clinically relevant events.

Diabetes affects ApoA-I-related outcomes

Prior research has indicated that HDL-C is associated with ischemic risk, but this association can be influenced by multiple factors, including glucose metabolism and blood pressure [16,26]. Given the rising prevalence of DM and the inconsistent findings regarding ApoA-I and cardiovascular risk, it is crucial to assess whether glucose metabolism status modulates ApoAI– related ischemic risk. Our results provide compelling evidence that the significant association between lower ApoA-I levels and increased ischemic risk is primarily observed in CAD patients with DM.

Two small-scale studies also suggested that the association between ApoA-I and ischemic risk appears to be stronger within DM subgroups. Drexel et al. [27] studied 491 patients with stable CAD receiving statin therapy and found that low ApoA-I levels were associated with a significantly increased risk of fatal or nonfatal cardiovascular events, especially in those with type 2 DM. Another study, which included 604 CAD patients undergoing PCI, showed that while ApoA-I levels were independently and negatively correlated with stent restenosis, combining ApoA-I and HbA1c levels improved the prediction of early stent restenosis (within 1 year), and combining ApoA-I with DM status enhanced prediction accuracy for late stent restenosis (beyond 1 year) [28]. These studies highlight the important association between ApoA-I and ischemic outcomes in conjunction with DM status in CAD patients; however, their limited sample sizes and short follow-up restrict generalizability. In contrast, our research included 10,232 CAD patients treated with PCI and followed them for up to 5 years, providing more robust data to support these findings. In the present study, we explicitly identified DM status as a key factor and demonstrated distinct clinical outcomes for ApoA-I according to different glucose metabolism statuses, particularly highlighting the impact of lower ApoA-I levels on long-term ischemic risk in patients with DM.

Potential mechanisms

The protective mechanism of ApoA-I in DM patients is not fully understood. Existing studies indicate that ApoA-I, as a major protein component of HDL particles, plays a critical role in cardiovascular protection. ApoA-I has been shown to stimulate cholesterol efflux and possesses anti-inflammatory and antioxidant properties [29]. Importantly, recent experimental research has further demonstrated that ApoA-I can suppress hyperglycemia-induced myelopoiesis, reduce circulating monocytes and neutrophils, and promote macrophage polarization toward an anti-inflammatory M2 phenotype. These effects contribute to atherosclerotic plaque regression in diabetic mice [30]. In this context, reduced ApoA-I levels may signal impaired HDL functionality and inadequate suppression of vascular inflammation, which could accelerate plaque progression and result in adverse cardiovascular outcomes. Although our study did not directly assess inflammatory pathways, the observed association between lower ApoA-I levels and increased MACCE risk in diabetic patients may be a consequence of these underlying mechanisms. Further research is needed to clarify and validate these potential pathways.

The lack of a significant association between lower ApoA-I levels and MACCE in individuals with pre-DM or NGR may be attributed to several factors. First, pathophysiological differences across the glycemic spectrum likely contribute. In diabetes, chronic hyperglycemia, oxidative stress, and inflammation are more pronounced, potentially amplifying the harmful effects of reduced ApoA-I. These processes are less prominent in pre-DM and NGR populations. Interestingly, our analyses also revealed that lower ApoA-I was associated with an increased risk of stroke in the pre-DM group, suggesting that early metabolic or vascular changes may already influence certain outcomes. In contrast, most outcomes in the NGR group showed no significant associations with ApoA-I, further supporting the idea that glucose metabolism status modifies the prognostic relevance of ApoA-I. Second, limited statistical power and the presence of competing risks may have contributed to the absence of significant associations in these groups. Larger studies are warranted to better define the role of ApoA-I in populations with less advanced glucose dysfunction.

Clinical implications and future directions

These results provide important insights for risk assessment and the identification of populations that may benefit from therapies targeting ApoA-I levels.

First, the present study underscores the significance of lower ApoA-I levels as a potential prognostic biomarker for identifying high-risk groups for ischemic events among CAD patients undergoing PCI. Importantly, our findings highlight the need to consider individual differences in glycemic metabolism status when using ApoA-I for risk stratification. Low ApoA-I levels alone may not indicate increased risk, but in combination with DM status, they may signify a higher likelihood of adverse outcomes. Our findings support the potential incorporation of ApoA-I measurements into routine risk stratification and monitoring protocols for CAD patients with DM undergoing PCI.

Furthermore, future research should explore therapies aimed at increasing ApoA-I levels to improve outcomes in diabetic patients. Previous pooled analyses have shown that apabetalone, a small molecule designed to upregulate endogenous ApoA-I synthesis, significantly reduced the risk of major adverse cardiovascular events in CAD patients with DM [31]. In addition, mimetic peptides developed to replicate the cardioprotective functions of ApoA-I are in early stages of development and have shown promising benefits. Various ApoA-I mimetic peptides, such as L-4F [32], D-4F [33], and RG541 [34], have demonstrated the ability to improve insulin sensitivity in animal models during preliminary experiments. Although research on ApoA-I as a therapeutic target is still in its early stages, our findings suggest that CAD patients with DM may be a key population to benefit from such interventions. Large-scale clinical trials are needed to determine whether strategies targeting ApoA-I levels can effectively reduce the risk of adverse events in CAD patients with DM.

Limitations

Several limitations should be acknowledged. First, this was a single-center observational study, which may limit the generalizability of our findings. Future validation in larger, multicenter populations is necessary. Second, lipid profiles and glucose levels may have fluctuated during the observational period, and various medications were administered, which could have influenced the outcomes. Future studies should incorporate dynamic monitoring of lipid and glucose levels to assess their impact on the results. Third, although this was a large-scale real-world study and adjustments were made for many relevant covariates, the observed associations may still be affected by unmeasured confounders, such as the use of oral antidiabetic medications. Fourth, although individual MACCE components were analyzed, some outcomes had relatively low incidence rates, potentially limiting the statistical power to detect significant associations. Further studies with larger sample sizes and longer follow-up durations are warranted to confirm these findings. Finally, while ApoA-I is a major component of HDL and associated with its functionality, it may not fully represent the functional properties of HDL. Future research incorporating direct assessments of HDL function is needed to clarify the relationship between HDL functionality and ischemic risk across different glucose metabolism statuses.

Conclusion

In this 5-year, large-scale cohort study, lower ApoA-I levels were associated with an elevated risk of MACCE in PCI patients with DM. These findings suggest that ApoA-I levels may serve as a valuable risk stratification indicator in CAD patients undergoing PCI with DM and may represent a future therapeutic target in this high-risk population.