2.1 Study Design and Participants

From January 2017 to December 2018, participants who underwent percutaneous coronary intervention (PCI) at Fuwai Hospital were consecutively enrolled, as illustrated in the flowchart (Figure 1). 1799 CAD participants with impaired kidney function (estimated glomerular filtration rate [eGFR] ≤ 60 mL/min/1.73 m², calculated using CKD-EPI equations [11]) undergoing PCI were included in the analysis cohort 1, and 13,578 CAD patients with DM were included in the analysis cohort 2, as illustrated in the flowchart. Exclusion criteria included incomplete data (such as baseline ET-1 and creatinine levels before angiography as well as diabetic status). Ultimately, 1344 CAD patients with impaired kidney function were included and categorized based on diabetes status. 10,577 DM participants were included and divided into three groups according to their baseline eGFR.

The study protocol was approved by the Ethics Committee of Fuwai Hospital, which was carried out in accordance with the Declaration of Helsinki's tenets (Approval Number 2016–847). All participants provided written, informed consent.

2.2 Clinical Status and Laboratory Techniques

The decision to perform PCI treatment and all PCI procedures was performed by skilled interventional cardiologists following established standardized guidelines and clinical considerations. Information regarding other comorbid conditions, family history, and previous treatments was obtained through self-reporting or hospital medical records.

Glycated hemoglobin (HbA1c) readings of ≥ 6.5%, fasting plasma glucose (FPG) levels of ≥ 7.0 mmol/L, or current usage of insulin or antihyperglycemic medications were used to diagnose diabetes. All blood samples were collected prior to coronary angiography and represented the first specimen obtained at admission, before any urgent administration of antiplatelet drugs, heparin, or other medications. A commercially available enzyme immunoassay (BI-20082H, Biomedica, Vienna, Austria) with a detection sensitivity of 0.02 pmol/L and a normal range of less than 0.25 pmol/L was used to measure the levels of circulating plasma big ET-1. Plasma glucose was determined using an enzyme-linked hexokinase assay, while HbA1c levels were assessed using an automated hemoglobin A1c analyzer (HLC-723G8, Tosoh, Tokyo, Japan).

2.3 Clinical Outcomes and Follow-Up

The study's main endpoint was major adverse cardiovascular events (MACE), which is a composite of all-cause mortality, cardiovascular mortality, non-fatal myocardial infarction, unplanned revascularization, and stroke. Death was classified as cardiac unless a clear non-cardiac explanation could be provided. Positive cardiac troponin levels along with typical chest pain, distinctive electrocardiogram abnormalities, angiographic evidence of intracoronary thrombus, or imaging evidence of acute myocardial necrosis or regional wall motion abnormalities were considered indicators of a non-fatal myocardial infarction. Neuroimaging investigations and thorough medical histories were used to diagnose stroke. Unplanned revascularization referred to any repeat PCI or surgical intervention occurring after hospital discharge, thus excluding scheduled staged PCI instances.

Patients were followed up every 6 months by trained investigators after discharge. Adverse event data were collected from medical records, clinical visits, and telephone interviews. Two qualified investigators, not involved in the study, independently assessed events using predefined criteria and a centralized review. For events with discrepancies, reassessment was performed by the third investigator.

2.4 Statistical Analysis

Categorical data were displayed as frequencies and percentages, whereas baseline characteristics for continuous variables were represented as medians with interquartile ranges (IQRs). The chi-square test was employed for categorical variables, while the Wilcoxon rank-sum test, Kruskal-Wallis test, or analysis of variance (ANOVA) were used to assess the statistical significance of differences for continuous data.

Because ET-1 pathways may be potentially affected in patients with renal impairment or DM, a more rigorous analysis was performed by setting different cutoff values for the two groups. Participants were split into three groups based on the tertiles of plasma ET-1 concentrations: low, intermediate, and high. The threshold values were 0.28 pmol/L and 0.44 pmol/L in individuals with impaired renal function, and 0.21 pmol/L and 0.33 pmol/L in those with DM. Kaplan–Meier survival analysis was employed to compare cumulative hazard rates. The relationships between ET-1 levels, diabetic status, kidney functions, and poor prognosis were analyzed through Cox regression analyses. The Cox proportional hazards assumption was tested via Schoenfeld Residuals. Model 1 was unadjusted, while Model 2 was a fully adjusted model for confounding factors including age, sex, body mass index (BMI), health status (acute coronary syndrome[ACS], smoking status, hypertension, dyslipidemia, family history of CAD), brain natriuretic peptide levels, medications at discharge (statins, aspirin, clopidogrel, ticagrelor, angiotensin II receptor blockers/angiotensin-converting enzyme inhibitors, beta-blockers, and calcium channel blockers), and angiographic characteristics (heavily calcified lesions, Syntax score, lesion length, chronic total occlusion, bifurcation, stent usage). A Fine-Gray subdistribution hazard model adjusted for competing risk as sensitivity analysis was conducted (Fine-Gray analysis for non-cardiac death as competing risk). Restricted cubic spline analysis was conducted to evaluate the presence of linear relationships between ET-1 and various adverse outcomes across different diabetes statuses and kidney functions.

A two-sided p-value of < 0.05 was considered statistically significant. All statistical analyses were performed using STATA version 17.0 and R version 4.1.3.

3.1 Baseline Characteristics

A total of 1344 CAD patients with renal dysfunction were included in this study; their median age was 69.8 years, and 59.4% of them were male. The baseline characteristics are summarized in Table 1 and Table S1. The median level of ET-1 was 0.35 pmol/L, with a mean of 0.47 pmol/L. Based on their ET-1 levels, patients were divided into three groups: low, intermediate, and high, with cut-off values of 0.28 and 0.44 pmol/L. Age, BNP, FPG, HbA1c, C-reactive protein (CRP), and creatinine levels all increased with rising plasma ET-1 concentrations. Additionally, the occurrence rates of ACS, atrial fibrillation (AF), and DM among CAD patients with poor kidney function also rose with increasing ET-1 levels. Conversely, left ventricular ejection fraction (LVEF), HDL, hemoglobin (HBG), and eGFR decreased as ET-1 levels increased. Depending on whether they had diabetes or not, the patients were split into two groups: those with diabetes and those without (Table S2). Notably, the ET-1 levels in diabetic patients were significantly higher than those in non-diabetic patients.

Table S3 summarized the baseline characteristics of the 10 577 CAD participants with DM in the study (median age: 61.4 years; male: 74.2%). Based on their eGFR, the individuals were divided into three groups as follows: eGFR ≥ 90 mL/min/1.73 m² (5274 subjects, 49.8%), 60 < eGFR < 90 mL/min/1.73 m² (4445 subjects, 42.0%), and eGFR ≤ 60 mL/min/1.73 m² (858 subjects, 8.2%). Notably, plasma levels of BNP, HbA1c, TG, TC, LDL, blood urea nitrogen (BUN), and ET-1 exhibited a positive correlation with baseline eGFR levels (all p < 0.05). Based on their ET-1 levels, patients were divided into three groups with cut-off values of 0.21 and 0.33 pmol/L: low, intermediate, and high.

3.2 ET-1 Levels and Clinical Outcomes

During the follow-up period (median follow-up time of 3.12 years, range 2.98 to 3.26 years), CAD patients with impaired kidney function underwent PCI, and 269 MACE occurred (20%), including 129 cases of all-cause mortality, 56 cases of cardiovascular mortality, 51 cases of non-fatal MI, 17 strokes, and 124 cases of unplanned revascularization therapy (Figure 2). Participants were also split into three groups according to plasma ET-1 concentrations with cut-off values of 0.28 and 0.44 pmol/L.

Univariable and multivariable Cox regression analysis indicated a significant association between elevated ET-1 levels and increased risk of MACE (Table 2). After full adjustment for relevant covariates, ET-1 level was positively associated with elevated risks of MACE (adjusted hazard ratio [HR], 1.333; 95% confidence interval [CI], 1.169–1.519, p < 0.001). Moreover, high ET-1 groups demonstrated a 1.695-fold increased risk of MACE compared with the low ET-1 group (adjusted HR, 1.694; 95% CI, 1.219–2.353, p < 0.001).

Furthermore, patients with elevated ET-1 levels experienced a markedly higher risk of cardiovascular mortality compared with those with lower ET-1 levels, consistent with the MACE results. In addition, compared with the reference group, those with elevated ET-1 levels had a noticeably higher risk of non-fatal myocardial infarction (adjusted HR, 2.203; 95% CI, 1.018–4.764, p < 0.001). However, no significant statistical association was found between ET-1 levels and stroke or unplanned revascularization (Table 2).

In CAD patients with DM who underwent PCI, 1363 participants experienced MACE, representing 12.9% of the cohort. This included 304 cases of all-cause mortality, 124 deaths due to cardiovascular causes, 282 nonfatal MI, 985 unplanned revascularizations, and 64 nonfatal strokes. Participants were categorized into three groups based on plasma ET-1 concentration tertiles: low ET-1 (≤ 0.21 pmol/L), medium ET-1 (0.21–0.33 pmol/L), and high ET-1 (> 0.33 pmol/L).

Elevated ET-1 levels were linked to a 1.492-fold increased risk of MACE (HR 1.492; 95% CI 1.311–1.698; p < 0.001), according to univariate Cox regression analysis (Table S4). In a multivariate Cox regression model adjusted for additional covariates, ET-1 was associated with MACE as a continuous variable (adjusted HR 1.149; 95% CI 1.057–1.248; p < 0.001). Further, elevated plasma ET-1 levels were significantly linked to MACE (adjusted HR 1.279; 95% CI 1.118–1.464; p < 0.001) compared with the low ET-1 group.

Elevated plasma ET-1 was associated with a higher risk of all-cause mortality (adjusted HR 2.188, 95% CI 1.585–3.022, p < 0.001) and cardiovascular mortality (adjusted HR 3.824, 95% CI 2.028–7.210, p < 0.001). A positive correlation between plasma ET-1 levels and the risk of nonfatal MI was also observed (for medium ET-1: adjusted HR 1.436; 95% CI 1.038–1.987, p < 0.001; for high ET-1: adjusted HR 1.970; 95% CI 1.436–2.702, p < 0.001). However, no statistically significant relationship was found between ET-1 levels and the incidence of strokes or unplanned revascularizations.

3.3 Diabetes, ET-1 Levels, and Clinical Outcomes

Participants with impaired kidney function were further categorized into non-diabetic (486, 36.1%) and diabetic groups (858, 63.9%). The risk of MACE varied between diabetic and non-diabetic patients with impaired kidney function, with diabetic patients exhibiting a higher incidence than their non-diabetic counterparts. Individuals with diabetes had a substantially higher cumulative risk of MACE than individuals without the disease, according to Kaplan–Meier curves. Furthermore, compared with patients with intermediate or low ET-1 levels, those with high ET-1 levels had the highest cumulative risk of MACE (Figure S1).

Crucially, when classifying patients with impaired kidney function according to their diabetes status and ET-1 levels, individuals with diabetes and intermediate to high ET-1 levels showed a markedly higher risk of MACE than the reference group, which consisted of those without diabetes and with low ET-1 levels (Figure 3). Diabetes patients with medium ET-1 levels and those with high ET-1 levels were found to have a higher risk of MACE by multivariable regression analysis, with an adjusted HR of 2.082 (95% CI, 1.291–3.357, p < 0.001) and 2.134 (95% CI, 1.334–3.413, p < 0.001), respectively (Table 3). The risk of MACE was higher in non-diabetic patients with high ET-1 levels than in those with low ET-1 levels; however, this link was not statistically significant when confounding variables were taken into account.

Additionally, there was no discernible difference in the risk of cardiovascular death between the reference group and non-diabetic patients with elevated ET-1 levels. Similarly, non-DM patients with high ET-1 levels did not significantly differ in their risk of cardiovascular death from the reference group, but diabetic patients with medium and high ET-1 levels had a statistically significant increased risk of cardiovascular mortality. Interestingly, the probability of a non-fatal myocardial infarction did not differ statistically significantly from the reference group across all combined groups (Tables S5–S7 and Figure S2).

3.4 Kidney Function, ET-1, and Clinical Outcomes

In patients with CAD and DM, MACE incidence rates differed by eGFR categories, recorded at 11.7%, 12.3%, and 23.4%, respectively (Figure 2). Among the groups, those with reduced kidney function had the lowest event-free survival rates (log-rank p < 0.001), according to Kaplan–Meier analysis. Furthermore, among the three groups, patients with the lowest event-free survival had the highest ET-1 levels (Figure S3).

Compared with patients with eGFR ≥ 90 mL/min/1.73 m² and low ET-1 levels, those with low eGFR and high ET-1 levels had a considerably higher risk of MACE (p < 0.001) (Figure 3). Multivariate Cox regression analyses showed that patients with eGFR ≤ 60 mL/min/1.73 m² and elevated ET-1 levels had a 2.297-fold (95% CI 1.822–2.895) increased risk of MACE, while those with medium ET-1 levels had a 1.638-fold (95% CI 1.179–2.275) increased risk. Patients with eGFR 60–90 mL/min/1.73 m² and elevated ET-1 levels had a 1.190-fold (95% CI 1.025–1.333) increased MACE risk, whereas those with eGFR ≥ 90 mL/min/1.73 m² and high ET-1 levels had a 1.356-fold (95% CI 1.117–1.646) increased risk (Table 4).

In terms of cardiovascular death and non-fatal myocardial infarction, patients with eGFR ≤ 60 mL/min/1.73 m² and intermediate to high levels of ET-1 have a significantly higher cumulative hazard of cardiovascular death and non-fatal myocardial infarction compared with other groups, showing statistical significance (p < 0.05) (Figure S4).

3.5 Non-Linear Relationships

The Restrictive Cubic Splines (RCS) analysis demonstrated a non-linear relationship between ET-1 and MACE in both diabetic and non-diabetic patients with CAD and impaired kidney function (P for nonlinear < 0.05) (Figure 4). Similarly, non-linear associations were observed between ET-1 and cardiovascular mortality and non-fatal myocardial infarction. In patients with CAD and DM, there is a non-linear relationship between the levels of ET-1 and MACE (P for nonlinearity < 0.05). This conclusion remains consistent across patients with varying levels of eGFR (Figure S5).

3.6 Competing Risk Analyses

The competing risk analyses were used with the multivariable Fine-Gray subdistribution hazard ratios (SHR) regarding non-cardiac death as a competing risk. In CAD patients with impaired kidney function, a higher risk for MACE was found in the high ET-1 level group (SHR 1.627, 95% CI 1.119–2.365, p = 0.011). Furthermore, diabetes patients with intermediate ET-1 levels and those with high ET-1 levels were found to have a higher risk of MACE with SHR of 1.961 (95% CI, 1.104–3.485, p = 0.022) and 2.192 (95% CI, 1.262–3.804, p = 0.005), respectively, which were consistent with Cox analysis.

In CAD patients with DM, ET-1 was also positively associated with MACE (SHR 1.610, 95% CI 1.474–1.759, p < 0.001). CAD patients with impaired kidney functions with intermediate and high ET-1 levels also had a higher MACE risk, with SHR of 3.509 (95% CI 1.841–6.689, p < 0.001) and 6.346 (95% CI 3.444–11.696, p < 0.001), respectively.