Integrated metabolic-inflammatory risk assessment: HbA1c/HDL-c and hsCRP-to-albumin ratios synergistically predict major adverse cardiovascular events in post-PCI STEMI patients

Abstract

Background

This study looked into how predictive in ST-segment elevation myocardial infarction (STEMI) is the high-sensitivity C-reactive protein-to-albumin ratio (hsCAR) as well as glycated hemoglobin to high-density lipoprotein cholesterol (HbA1c/HDL-c) ratio.

Methods

This retrospective cohort research was carried out on 1,177 patients having STEMI who were given percutaneous coronary intervention (PCI). For major adverse cardiovascular events (MACE), the independent and combined predictive values of the HbA1c/HDL-c ratio (threshold ≥ 6.61) and hsCAR (threshold ≥ 0.18) were assessed. MACE was an amalgamation of death from all causes, ischemia-induced revascularization, myocardial infarction not leading to death, heart failure hospitalization, and cerebrovascular events. The team used Cox regression models, causal mediation examination, and receiver operating characteristic curves to assess prognostic performance and mechanistic pathways, and compared them with the Global Registry of Acute Coronary Events (GRACE) risk score.

Results

The interquartile range for follow-up was 79 to 672 days, with 461 median days. A raised HbA1c/HDL-c ratio (≥ 6.61) and hsCAR (≥ 0.18) independently predicted MACE, with 1.51 (95% confidence interval [CI]: 1.26–1.81; P < 0.001) as well as 1.84 (95% CI: 1.53–2.21; P = 0.005) hazard ratios (HRs), respectively. Combined use enhanced risk stratification, with the high HbA1c/HDL-c–high hsCAR group showing the highest risk (adjusted HR 2.19, 95% CI: 1.67–2.87; P < 0.001). Causal mediation examination revealed that coronary lesion complexity partially mediated these associations; the SYNergy between PCI with TAXUS and Cardiac Surgery (SYNTAX) and the residual SYNTAX scores were responsible for accounting for 24.2%/17.7% of the hsCAR effect and 16.8%/25.2% of the HbA1c/HDL-c ratio effect, respectively. Compared with the individual markers or the GRACE risk score, the combined biomarker model demonstrated superior discriminatory capacity (area under the curve = 0.63, 95% CI = 0.60–0.66; P < 0.001), with significant improvement in integrated discrimination.

Conclusion

The integration of HbA1c/HDL-c and hsCAR can significantly improve risk stratification in patients with STEMI, outperforming traditional scoring systems and assisting in the precise management of individuals at risk.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12944-025-02707-2.

Introduction

One of the main reasons of morbidity and death throughout the world today is coronary heart disease. Even after advancements in interventional techniques as well as pharmacological therapies that have substantially improved early survival rates in ST-segment elevation myocardial infarction (STEMI), these individuals face considerable risk of long-term cardiovascular problems after percutaneous coronary intervention (PCI) [1]. Consequently, identifying more precise risk assessment tools, particularly biomarker combinations that integrate multiple pathophysiological mechanisms, holds significant clinical importance for improving outcomes in patients with STEMI.

STEMI pathogenesis covers a complex interplay between disruptions in the body’s normal metabolic processes and inflammatory processes, which are intimately interconnected and mutually reinforcing [2]– [3]. In the acute phase of myocardial infarction, metabolic perturbations, including insulin resistance, hyperglycemia, and dyslipidemia, create a proatherogenic environment that accelerates plaque instability and promotes thrombosis [4]– [5]. Simultaneously, the inflammatory cascade triggered by myocardial necrosis leads to systemic inflammatory activation, which not only exacerbates tissue damage but also influences long-term cardiovascular risk through effects on endothelial function, plaque vulnerability, and coronary microcirculation [2, 6]. While individual biomarkers provide valuable insights into specific pathophysiological pathways, emerging evidence suggests that composite markers integrating multiple parameters may offer superior predictive capacity. This integrated approach better captures the multifaceted nature of cardiovascular disease progression, potentially enabling more precise risk stratification and personalized therapeutic strategies.

In 2021, Hu et al. proposed the glycated hemoglobin to high-density lipoprotein cholesterol (HbA1c/HDL-c) ratio, which has since become a likely metabolic composite marker [7]. This ratio uniquely integrates two crucial aspects of cardiometabolic health: long-term glycemic control (via HbA1c, showing the mean glucose levels during 2 to 3 months) and lipid metabolism functionality (via HDL-c, reflecting cholesterol reverse transport and anti-inflammatory capacity). By simultaneously capturing these complementary dimensions, the HbA1c/HDL-c ratio provides a more comprehensive assessment of metabolic dysregulation than either parameter alone.

Since its introduction, the HbA1c/HDL-c ratio shows clinical utility across multiple conditions, including carotid atherosclerosis [7] and stroke risk estimation [8]. These investigations indicate that this ratio not only reflects insulin resistance states but also captures the composite risk arising from prolonged hyperglycemia-induced vascular endothelial damage coupled with diminished HDL-c protective function. Nevertheless, no research has thoroughly examined the link between major adverse cardiovascular events (MACE) risk and the HbA1c/HDL-c ratio in STEMI after PCI, despite these encouraging uses.

Complementing metabolic assessments, inflammatory responses govern atherosclerosis as well as myocardial infarction. A high-sensitivity C-reactive protein (hsCRP)-to-albumin ratio (hsCAR) represents an inflammatory indicator that has demonstrated prognostic value in various cardiovascular scenarios [9–13]. By combining a proinflammatory marker (hsCRP) with a parameter reflecting nutritional status and anti-inflammatory capacity (albumin), this ratio offers additional insights into systemic inflammatory status that may complement metabolic risk assessment.

One method that shows promise for risk stratification in STEMI is the combination of metabolic and inflammatory biomarkers, given the well-established interaction between these systems during acute myocardial infarction [14]– [15]. Metabolic dysfunction promotes inflammation through advanced glycation, oxidative stress, and dyslipidemia [16], whereas inflammation can worsen metabolic disturbances via cytokine-induced insulin resistance and stress hyperglycemia [17]– [18]. This two-way interaction may amplify cardiovascular risk beyond the contribution of either factor alone.

Both processes influence important post-STEMI outcomes, including coronary complexity, microvascular function, ventricular remodeling, and thrombosis risk [19]. However, the combined prognostic value of the HbA1c/HDL-c ratio (showing a chronic metabolic condition) and inflammatory markers remains underexplored in PCI-treated patients with STEMI. This gap is important to address, given patient heterogeneity in metabolic and inflammatory status, which may affect long-term outcomes.

Based on this background, this study looked into the independent and combined predictive power of the hsCAR alongside the HbA1c/HDL-c ratio after PCI for MACE in STEMI. The hypothesis is that the HbA1c/HDL-c ratio provides noteworthy prognostic factors beyond conventional risk factors, and integrating metabolic and inflammatory assessments further enhances risk stratification by capturing the synergistic effects of these interconnected pathophysiological pathways, potentially informing tailored management strategies for this high-risk population.

Methods

Study design and participants

A comprehensive retrospective analysis was carried out on consecutive people having STEMI enrolled at the Tianjin Medical University General Hospital’s Cardiovascular Disease Registry from August 2018 to December 2023. Tianjin Medical University General Hospital’s Ethics Committee approved the protocol (approval number: IRB2023-YX-301-01/2023), which matched the philosophies mentioned in the Declaration of Helsinki (1975) and its subsequent revisions. Taking into account the retrospective data analysis based on standard clinical records, written informed consent was not needed.

The research team selected people aged ≥ 18 years who had the following conditions for a definitive STEMI diagnosis [20]: (i) characteristic ischemic symptoms lasting ≥ 30 min; (ii) = ST-segment rise ≥ 1 mm in electrocardiography by a new left bundle branch block or minimum two adjoining leads; and (iii) increased cardiac troponin exceeding the 99th reference percentile. Researchers eliminated 303 of the 1,480 consecutive patients based on these predefined criteria: (1) no coronary angiography or angiography without treatment (n = 33); (2) liver impairment (aspartate aminotransferase or alanine aminotransferase > thrice the upper normal limit), advanced renal damage (< 30 mL/min/1.73 m² estimated glomerular filtration rate), or simultaneous active infection (n = 23); (3) PCI, coronary artery bypass grafting (CABG), or myocardial infarction (n = 140), which aims to reduce confounding from chronic disease status and prior interventions, thereby allowing a more accurate assessment of baseline biomarker associations with post-PCI outcomes; (4) insufficient clinical or laboratory data pertinent to glycemic parameters or lipid profiles (n = 31); and (5) inability to complete the follow-up protocol (n = 76). The final analytical cohort comprised 1,177 patients who underwent a standardized evaluation and treatment protocol (Supplementary Fig. S1).

Baseline demographic and clinical characteristics were systematically recorded at admission. Current smoking status was defined as the consumption of ≥ 1 cigarette daily within 30 days preceding hospital admission [21]. Diabetes mellitus was defined by either an established prehospitalization diagnosis or documented use of glucose-lowering agents. One of these events meant hypertension: (1) previous clinical diagnosis, (2) antihypertensive medication use prior to admission, or (3) new diagnosis during the index hospitalization based on repetitive blood pressure records beyond 140/90 mmHg.

Sample size

Researchers calculated the study sample based on Cox proportional hazards model assumptions to detect a 1.5 hazard ratio with 80% statistical power at a two-sided 0.05 value showing significance. Drawing upon previously published data on cardiovascular event rates among patients with STEMI [22]– [23], a 15% cumulative incidence of MACE was estimated over 2 years of follow-up. These parameters yielded the required sample size of 637 participants. After applying a 10% adjustment factor for potential losses to follow-up, the minimum required enrollment was established at 708 patients. The final cohort of 1,177 participants substantially exceeded this requirement, ensuring robust statistical power for both primary analyses and prespecified subgroup investigations.

Comprehensive data collection and laboratory assessment

Researchers used electronic medical records alongside a standardized data collection protocol to extract comprehensive information. Demographic figures included sex, anthropometric measurements (height and weight, which helped us compute the body mass index [BMI]), age, smoking pattern, and vital parameters (heart rate and blood pressure). The medical history included diagnoses of diabetes mellitus, hypertension, and cerebrovascular events.

Blood samples were collected immediately upon admission before administration of antiplatelet agents or anticoagulants. Certified laboratory technicians, unaware of the clinical data, carried out laboratory assessments using standardized assays on automated platforms. Hematological parameters included complete blood counts with differential, comprehensive metabolic panels, and specialized cardiovascular biomarkers. Specifically, platelet count, hemoglobin concentration, serum albumin, renal function parameters (serum creatinine + eGFR obtained via the Modification of Diet in Renal Disease formula), cardiac biomarkers (B-type natriuretic peptide [BNP]), glycemic indices (HbA1c as well as fasting blood glucose [FBG]), and lipid profile (triglycerides [TG], total cholesterol [TC], HDL-c, and low-density lipoprotein cholesterol [LDL-c]) were quantified.

The team used Beckman Coulter, Inc.’s latex-enhanced immunoturbidimetric test to find HsCRP, with a 0.02 mg/L detection limit alongside inter- and intra-test coefficients of variation < 5%. Researchers found HbA1c using Tosoh G8 high-performance liquid chromatography (Tosoh Bioscience, Inc.) with standardization according to the National Glycohemoglobin Standardization Program. HDL-c was quantified using a direct enzymatic colorimetric assay following selective precipitation of apolipoprotein B-containing lipoproteins. See Supplementary Table S1 for details.

From these laboratory parameters, two composite indices were calculated:

• HbA1c/HDL-c ratio : HbA1c percentage ÷ HDL-cholesterol concentration (mmol/L).

• hsCAR : hsCRP (mg/L) ÷ serum albumin concentration (g/L).

Cardiac function was assessed via transthoracic echocardiography performed within a day of admission by veteran cardiologists by means of standardized protocols. The Biplane Simpson’s process helped us measure the left ventricular ejection fraction (LVEF).

Coronary intervention procedure and angiographic assessment

All patients were given PCI after coronary angiography per current technical guidelines [24]. A loading dose containing 300 mg acetylsalicylic acid and either 180 mg ticagrelor or 300 mg clopidogrel were given before the procedure, based on clinical signs and bleeding risk test. Unless otherwise mentioned, antiplatelet therapy after the procedure involved taking 100 mg acetylsalicylic acid every day alongside either 90 mg ticagrelor two times a day or 75 mg clopidogrel regularly for at least a year.

Two skilled interventional cardiologists who were unaware of clinical information used the SYNergy between PCI with TAXUS and the Cardiac Surgery (SYNTAX) scoring system to systematically assess the complexity of coronary lesions. When scores differed, a third cardiologist made a decision to reach an agreement. After PCI, the team calculated the untreated coronary disease burden via the residual SYNTAX score (rSS). Both the initial SYNTAX score and the rSS have demonstrated prognostic utility in previous investigations [25].

All patients were given standard evidence-based pharmacotherapy, consisting of statins, beta-blockers, angiotensin receptor blockers [ARBs], renin-angiotensin-aldosterone system inhibitors (angiotensin-converting enzyme inhibitors [ACEIs], or angiotensin receptor-neprilysin inhibitors [ARNIs]). Additionally, researchers prescribed proprotein convertase subtilisin/kexin type 9 inhibitors (PCSK9is) as well as sodium-glucose cotransporter 2 inhibitors (SGLT2is) per specific clinical indications, glycemic parameters, cardiac function assessment, and lipid profiles in accordance with contemporary guidelines.

Follow-up protocol and endpoint definitions

In this retrospective study, follow-up data were systematically obtained after checking electronic medical records from both outpatient visits and subsequent hospitalizations. To ensure the complete ascertainment of events, trained research personnel conducted structured telephone interviews using standardized questionnaires for patients with incomplete documentation in the electronic health record system. The combination of nonfatal myocardial infarction (per the Fourth Universal Definition [20]), death from all causes, ischemia-caused repeat revascularization, cerebrovascular complications (transient ischemic attack or ischemic stroke established by imaging alongside neurological tests), or heart failure hospitalization (needing inotropes, intravenous diuretics, or vasodilators) comprised MACE, which was the primary endpoint. To maintain rigor in this retrospective analysis, standardized criteria were applied consistently across all event evaluations, with particular attention to distinguishing between index hospitalization complications and true post-discharge events. Researchers computed the time-to-event from the index PCI date to the first established MACE date, with censoring at the date of the last documented follow-up or study conclusion for event-free patients.

Statistical analysis methodology

After analyzing for normality, researchers use the means ± standard deviations to show continuous variables, or, if appropriate, medians with interquartile ranges. The team uses percentages and frequencies to show categorical variables. For continuous variables, researchers carried out the t-test/Mann-Whitney U test to find between-group variances; for categorical variables, they carried out Fisher’s exact/chi-square tests. The team created a directed acyclic plot to show the associations among hsCAR, MACE, and the HbA1c/HDL-c ratio (Supplementary Fig. S2).

Researchers found associations among MACE, hsCAR, and the HbA1c/HDL-c ratio by creating multivariable Cox proportional hazards models. Nonlinear associations were assessed using restricted cubic spline (RCS) functions. Cutoff values for HbA1c/HDL-c and hsCAR were determined based on optimal thresholds identified from log-rank statistics and spline inflection points, respectively.

Based on the established threshold values, the team stratified patients as: (1) lowHbA1c/HDL-c ratio with low hsCAR; (2) low HbA1c/HDL-c ratio with high hsCAR; (3) high HbA1c/HDL-c ratio with low hsCAR; and (4) high HbA1c/HDL-c ratio with high hsCAR. The Kaplan-Meier technique helped us create cumulative incidence curves, and the log-rank test helped us find variances between groups.

To evaluate HbA1c/HDL-c ratio’s and hsCAR’s prognostic value (together as well as alone), a series of hierarchical Cox proportional hazards models were developed via progressive modification for confounding variables: Model 1: not adjusted; Model 2: for demographic factors (sex, BMI, and age); Model 3: Model 2 alongside clinical indices (blood pressure, heart rate, smoking pattern, diabetes, and stroke) and LVEF; Model 4: Model 3 alongside pharmacotherapy (antiplatelet regimen, ARBs/ARNIs/ACEIs, and beta-blockers) and laboratory parameters (platelet count, hemoglobin, eGFR, LDL-c); and Model 5: application of bootstrapping techniques with 1,000 resampling iterations to enhance the robustness of effect estimates. Researchers affirmed the proportional hazards assumption via log(-log) survival plots as well as scaled Schoenfeld residuals. Values less than three were deemed allowable when checking multicollinearity through variance inflation factors.

To elucidate potential mechanistic pathways linking the HbA1c/HDL-c ratio and hsCAR to MACE, formal causal mediation analyses were carried out using the approach defined by Baron and Kenny, which was extended with bootstrapping procedures [26]. This methodology allows for decomposition of the total effect of each biomarker into direct effects and indirect effects mediated through coronary lesion burden (assessed via the SYNTAX score and rSS). Researchers carried out a causal mediation test to evaluate if the coronary lesion burden governs HbA1c/HDL-c’s and hsCAR’s prognostic effects.

Subgroup analyses explored potential differences in treatment effects according to age, sex, SYNTAX score category, diabetes status, and SGLT2i/PCSK9i therapy. Researchers modified Cox models to add interaction terms and test for treatment effect heterogeneity across subgroups. Hazard ratios with interaction P values alongside 95% confidence intervals (CIs) are shown in forest plots.

The team calculated the area under the curve (AUC) with a 95% CI was computed using the receiver operating characteristic (ROC) curve, to find the HbA1c/HDL-c ratio’s and hsCAR’s incremental prognostic value. Improvements in risk discrimination and reclassification over existing models (e.g., the Global Registry of Acute Coronary Events [GRACE] and modified GRACE risk scores) were quantified using category-free net reclassification improvement (NRI) as well as integrated discrimination improvement (IDI) [27]. Likelihood ratio tests helped us find improvements in model fit.

Considering the established link between the MACE and triglyceride glucose (TyG) index, researchers carried out a supplemental test to evaluate the TyG index’s effect on MACE and the predictive capability of the TyG index along with the hsCAR.

To assess whether these losses were random or systematic, they compared the starting point features of all patients, and using inverse probability weighting, the team carried out a sensitivity analysis to evaluate how potential bias affected the results.

R 4.3.0 (R Foundation for Statistical Computing, Vienna, Austria) helped us with all tests. Significant results were described as two-sided P values less than 0.05. The Benjamini-Hochberg technique helped us regulate the false discovery rate during multiple comparisons.

Results

Study population and baseline characteristics

Between August 2018 and December 2023, researchers carried out this retrospective study at Tianjin Medical University General Hospital, and initially, 1,480 patients with STEMI were identified. Because they were unable to finish the follow-up owing to the loss of contact (e.g., invalid contact information or relocation), they excluded 76 patients. Finally, 1,177 patients were enrolled. During 461 days of median follow-up (interquartile range: 79–672 days), 483 patients (41.0%) experienced MACE.

When the team compared the starting point features of patients with and without MACE, they found a number of noteworthy variances (Supplementary Table S2). Patients who developed MACE were characterized by advanced age, more complex coronary artery disease (elevated SYNTAX scores and rSS), and more diabetes and hypertension. The MACE group also demonstrated compromised renal function (reduced eGFR), enhanced systemic inflammation (elevated hs-CRP and hsCAR), and adverse metabolic profiles (increased HbA1c, FBG, and HbA1c/HDL-c ratio). Additionally, these patients more frequently required multiple stent implantation and exhibited reduced LVEF.

Prognostic value of the HsCAR and HbA1c/HDL-c ratio for maces

The HbA1c/HDL-c ratio and hsCAR frequencies are illustrated in Supplementary Fig. S3. Multivariable RCS analysis revealed different relationship patterns between these biomarkers and MACE risk (Fig. 1A-B). The HbA1c/HDL-c ratio was linearly related to MACE risk, with maximally selected rank statistics identifying an optimal cutoff value of 6.61. In contrast, hsCAR exhibited a nonlinear relationship with MACE risk, with a significance threshold of hsCAR ≥ 0.18, beyond which the hazard ratio increased substantially.

Fig. 1.

Relationship between metabolic-inflammatory biomarkers and MACE risk in patients with STEMI. A Multivariable restricted cubic spline test demonstrating the predominantly linear association of the HbA1c/HDL-c ratio with MACE (optimal cutoff: 6.61, determined by maximally selected rank statistics). B Multivariable restricted cubic spline test shows the nonlinear link of hsCAR with MACE risk (inflection point: 0.18). C Kaplan-Meier curves stratified by the HbA1c/HDL-c ratio (<6.61 vs. ≥6.61), showing significantly divergent survival rates without MACE (HR 1.51, 95% CI: 1.26‒1.81; P<0.001).D Kaplan-Meier curves stratified by hsCAR (<0.18 vs. ≥0.18), illustrating substantial differences in survival without MACE (HR 1.84, 95% CI: 1.53–2.21; P=0.005). Shaded areas around the survival curves represent 95% confidence intervals.

Kaplan-Meier curves with conforming log-rank test outcomes (Fig. 3C-D) illustrated the differential MACE risk stratified by these biomarkers according to their respective cutoff points. Individuals with an elevated HbA1c/HDL-c ratio (≥ 6.61) presented a greater risk of MACE (< 6.61) (hazard ratio [HR] 1.51, 95% CI: 1.26–1.81; P < 0.001). Similarly, the high- hsCAR group (≥ 0.18) demonstrated a more pronounced risk of MACE (HR 1.84, 95% CI: 1.53–2.21; P < 0.001) than those with lower hsCAR values, suggesting that compared with the HbA1c/HDL-c ratio, hsCAR might be a more potent indicator of adversative cardiovascular consequences.

Fig. 3.

Mechanistic pathways linking biomarkers to cardiovascular outcomes: causal mediation analysis. Schematic representation of direct and indirect effects through coronary lesion complexity. For the HbA1c/HDL-c ratio, the residual SYNTAX score (rSS) alongside the SYNTAX score mediated 25.2% and 16.8% of the total effect, respectively. For hsCAR,

Combined prognostic value of the HsCAR and HbA1c/HDL-c ratio

Based on the established cutoff points, researchers divided the population into: (1) low HbA1c/HDL-c ratio (< 6.61) with low hsCAR (< 0.18); (2) low HbA1c/HDL-c ratio with high hsCAR (≥ 0.18); (3) high HbA1c/HDL-c ratio (≥ 6.61) having low hsCAR; and (4) high HbA1c/HDL-c ratio with high hsCAR. Table 1 shows their features at the starting point.

Table 1.

Comparison of baseline characteristics by HbA1c/HDL-c ratio and HsCAR grouping. Median (interquartile range) shows continuous variable data, and numbers (percentages) show categorical variable data. Researchers stratified patients into four classes based on HbA1c/HDL-c ratio (< 6.61 vs. ≥ 6.61) and hsCAR levels (< 0.18 vs. ≥ 0.18 mg/L). Differences between continuous variable groups were found via the Kruskal-Wallis test, but for categorical variable groups via Fisher’s exact test/the chi-square test. P values represent overall comparisons across all four groups

	Total n = 1177	HbA1c/HDL-c < 6.61		HbA1c/HDL-c ≥ 6.61		P value
		hsCAR < 0.18 n = 384	hsCAR ≥ 0.18 n = 301	hsCAR < 0.18 n = 222	hsCAR ≥ 0.18 n = 270
Age (years)	66 (57, 72)	65 (55, 70)	68 (59, 74)	64 (54, 71)	66 (57, 72)	< 0.001
Female, n(%)	265 (23%)	92 (24%)	72 (24%)	43 (19%)	58 (21%)	0.50
BMI (kg/㎡)	24.8 (23.1, 27.5)	24.5 (22.8, 26.6)	24.5 (22.9, 27.0)	25.6 (23.6, 28.4)	25.7 (23.4, 28.2)	< 0.001
Heart rate (bpm)	78 (69, 90)	76 (68, 87)	80 (70, 92)	76 (66, 84)	81 (69, 94)	< 0.001
SBP(mmHg)	137 (121, 154)	139 (123, 153)	138 (124, 153)	140 (122, 156)	132 (116, 152)	0.009
DBP (mmHg)	84 (74, 95)	86 (77, 95)	85 (73, 95)	85 (75, 96)	81 (70, 93)	0.013
Current smoking, n(%)	549 (47%)	169 (44%)	135 (45%)	108 (49%)	137 (51%)	0.30
Hypertension, n(%)	798 (68%)	250 (65%)	200 (66%)	160 (72%)	188 (70%)	0.30
Diabetes, n(%)	348 (30%)	53 (14%)	48 (16%)	103 (46%)	144 (53%)	< 0.001
Stroke, n(%)	159 (14%)	42 (11%)	46 (15%)	31 (14%)	40 (15%)	0.30
Interventions
Stent, n(%)						0.003
0	105 (8.9%)	24 (6.3%)	28 (9.3%)	24 (11%)	29 (11%)
1	682 (58%)	255 (66%)	164 (54%)	126 (57%)	137 (51%)
≥ 2	390 (33%)	105 (27%)	109 (36%)	72 (32%)	104 (39%)
SYNTAX score	19 (13, 24)	17 (11, 22)	20 (15, 26)	19 (13, 24)	20 (13, 26)	< 0.001
rSS	7 (2, 11)	5 (1, 10)	7 (3, 12)	7 (4, 12)	8 (3, 13)	< 0.001
GRACE score	143 (123, 162)	137 (119, 153)	150 (131, 167)	134 (115, 155)	147 (129, 165)	< 0.001
LVEF (%)	48 (42, 55)	48 (43, 55)	45 (40, 50)	50 (44, 57)	46 (40, 52)	< 0.001
Platelet (*109/L)	222 (188, 264)	219 (190, 260)	229 (190, 266)	218 (183, 262)	225 (182, 265)	0.50
Hemoglobin (g/L)	144 (130, 156)	146 (132, 157)	142 (128, 154)	147 (133, 156)	140 (127, 154)	0.011
eGFR (ml/min/1.73㎡)	97 (79, 115)	99 (85, 114)	96 (78, 114)	97 (77, 115)	93 (70, 117)	0.030
Serum albumin (g/L)	39.0 (35.0, 43.0)	41.0 (37.0, 44.0)	38.0 (34.0, 42.0)	40.0 (37.0, 44.0)	37.0 (33.0, 41.0)	< 0.001
FBG (mmol/L)	6.30 (5.30, 8.30)	5.80 (5.10, 6.70)	5.80 (5.20, 7.10)	7.20 (5.60, 10.00)	7.85 (6.03, 11.00)	< 0.001
HbA1C (%)	6.10 (5.60, 7.20)	5.80 (5.50, 6.20)	5.80 (5.50, 6.20)	7.00 (6.13, 8.50)	7.40 (6.20, 9.00)	< 0.001
hs-CRP (mg/L)	7 (3, 15)	3 (1, 4)	15 (10, 31)	3 (2, 5)	15 (11, 35)	< 0.001
BNP (pg/mL)	75 (24, 264)	50 (17, 114)	178 (49, 394)	51 (19, 107)	151 (30, 515)	< 0.001
TC (mmol/L)	4.65 (3.97, 5.38)	4.89 (4.16, 5.53)	4.83 (4.09, 5.54)	4.37 (3.71, 5.06)	4.36 (3.71, 5.11)	< 0.001
TG (mmol/L)	1.58 (1.22, 2.20)	1.44 (1.07, 1.95)	1.47 (1.15, 1.95)	1.80 (1.35, 2.57)	1.77 (1.36, 2.49)	< 0.001
LDL-c(mmol/L)	2.96 (2.41, 3.59)	3.06 (2.52, 3.74)	3.04 (2.40, 3.67)	2.83 (2.24, 3.37)	2.85 (2.30, 3.50)	< 0.001
HDL-c (mmol/L)	1.01 (0.87, 1.17)	1.12 (1.01, 1.27)	1.12 (0.99, 1.25)	0.87 (0.78, 0.98)	0.86 (0.77, 0.97)	< 0.001
HbA1c/HDL-c	6.26 (5.19, 7.71)	5.30 (4.67, 5.93)	5.47 (4.75, 6.02)	8.03 (7.11, 9.49)	8.38 (7.37, 10.26)	< 0.001
hsCAR (mg/L)	0.17 (0.07, 0.39)	0.07 (0.03, 0.11)	0.40 (0.27, 0.83)	0.09 (0.05, 0.13)	0.40 (0.26, 0.93)	< 0.001
P2Y12i, n(%)						0.053
Clopidogrel	446 (38%)	141 (37%)	127 (42%)	69 (31%)	109 (40%)
Ticagrelor	731 (62%)	243 (63%)	174 (58%)	153 (69%)	161 (60%)
Statin, n(%)						0.40
Rosuvastatin	1,056 (90%)	341 (89%)	275 (91%)	203 (91%)	237 (88%)
Atorvastatin	121 (10%)	43 (11%)	26 (8.6%)	19 (8.6%)	33 (12%)
ACEI/ARB/ARNI, n(%)	383 (33%)	106 (28%)	103 (34%)	84 (38%)	90 (33%)	0.06
Beta blocker, n(%)	702 (60%)	211 (55%)	182 (60%)	138 (62%)	171 (63%)	0.13
PCSK9i, n(%)	185 (16%)	59 (15%)	51 (17%)	35 (16%)	40 (15%)	0.90
SGLT2i, n(%)	227 (19%)	28 (7.3%)	38 (13%)	68 (31%)	93 (34%)	< 0.001
Follow time (day)	461 (79, 672)	576 (340, 765)	426 (57, 646)	444 (76, 629)	375 (48, 596)	< 0.001
MACE, n(%)	483 (41%)	104 (27%)	139 (46%)	86 (39%)	154 (57%)	< 0.001

BMI Body mass index, SBP Systolic blood pressure, DBP Diastolic blood pressure, SYNTAX score SYNergy between PCI with TAXUS and the cardiac surgery score, rSS Residual SYNTAX score, LVEF Left ventricular ejection fraction, FBG Fasting blood glucose, hs-CRP High-sensitivity C-reactive protein, BNP B-type natriuretic peptide, TC Total cholesterol, TG Triglyceride, LDL-c Low-density lipoprotein cholesterol, HDL-c High-density lipoprotein cholesterol, hsCAR High-sensitivity C-reactive protein-to-albumin ratio, P2Y12i P2Y12 Receptor Inhibitors, ACEI Angiotensin-converting enzyme inhibitor, ARB Angiotensin receptor blocker, ARNI Angiotensin receptor and neprilysin inhibitor, PCSK9i PCSK9 inhibitors, SGLT2i Sodium–glucose cotransporter 2 inhibitor, MACE Adverse cardiovascular events

People in the low HbA1c/HDL-c ratio -low hsCAR group showed a lower diabetes prevalence than those in groups with minimum one raised biomarker. Additionally, these patients exhibited progressively elevated SYNTAX scores and rSSs, increased indices of glucose metabolism and inflammatory markers, and a greater proportion received SGLT2i therapy following hospitalization.

Table 2; Fig. 2 demonstrate the synergistic effect of the combined HbA1c/HDL-c ratio and hsCAR on MACE incidence. The MACE risk was significantly higher in the three groups with minimum one raised biomarker in the Model 1 than the low HbA1c/HDL-c ratio -low hsCAR group. Following sequential adjustments for demographic characteristics, clinical parameters, and angiographic variables, the elevated risk persisted in both the high hsCAR -low HbA1c/HDL-c ratio group (HR 1.88, 95% CI = 1.45–2.44) and the high hsCAR -high HbA1c/HDL-c ratio group (HR 2.19, 95% CI = 1.67–2.87) in the fully adjusted model (Model 4). The robustness of these findings was further validated through bootstrap analysis with 1,000 replicates, with the associations maintaining statistical significance across iterations.

Table 2.

Impact of HbA1c/HDL-c ratio combined with HsCAR on MACE. Hazard ratios (95% confidence intervals) are presented for MACE occurrence across the four stratified groups; the low hsCAR group alongside low HbA1c/HDL-c ratio are as the reference category. Model 1: not adjusted analysis. Model 2: demographic factors (age, sex, BMI) + Model (1) Model 3: clinical parameters (blood pressure, heart rate, smoking pattern, diabetes, stroke) and LVEF + Model (2) Model 4: building upon Model 3 with additional adjustment for pharmacotherapy (beta-blockers, antiplatelet regimen, ARBs/ARNIs/ACEIs), and laboratory parameters (platelet count, hemoglobin, LDL-c, and eGFR). Model 5: application of bootstrapping techniques with 1,000 resampling iterations to enhance the robustness of effect estimates. P for trend was calculated by treating the four groups as an ordinal variable. Researchers used Cox proportional hazards regression. *P < 0.05, **P < 0.01, ***P < 0.001. Ref., reference

	HbA1c/HDL-c < 6.61		HbA1c/HDL-c ≥ 6.61		P for trend
	hsCAR < 0.18 n = 384	hsCAR ≥ 0.18 n = 301	hsCAR < 0.18 n = 222	hsCAR ≥ 0.18 n = 270
Model 1	Ref.	1.97(1.53–2.54)***	1.62(1.22–2.16)***	2.52(1.97–3.24)***	< 0.001
Model 2	Ref.	1.89(1.46–2.44)***	1.63(1.22–2.17)***	2.45(1.90–3.15)***	< 0.001
Model 3	Ref.	1.85(1.54–2.40)***	1.50(1.12–2.02)**	2.10(1.61–2.74)***	< 0.001
Model 4	Ref.	1.88(1.45–2.44)***	1.60(1.18–2.16)**	2.19(1.67–2.87)***	< 0.001
Model 5	Ref.	1.88(1.43–2.47)***	1.60(1.16–2.21)**	2.19(1.66–2.90)***	< 0.001

Model 1 is considered the unadjusted model

Model 2 adjusts for sex, age, and BMI

Model 3 includes the variables in Model 2 with the addition of heart rate, SBP, DBP, current smoking status, diabetes status, stroke status, and LVEF

Model 4 builds upon Model 3 by further adjusting for antiplatelet therapy, beta-blockers, ACEI/ARB/ARNI, hemoglobin, platelet count, eGFR, and LDL-c

Model 5 is an extension of Model 4 with the addition of bootstrapping for statistical robustness

* denotes P < 0.05

** denotes P < 0.01

*** denotes P < 0.001

Fig. 2.

Synergistic effect of hsCAR stratification plus the HbA1c/HDL-c ratio on MACE incidence. K-M curves for four patient subgroups: low HbA1c/HDL-c-low hsCAR (reference), low HbA1c/HDL-c-high hsCAR (HR 1.62, 95% CI: 1.25–2.11), high HbA1c/HDL-c-low hsCAR (HR 1.38, 95% CI: 1.03–1.85), and high HbA1c/HDL-c-high hsCAR (HR 1.94, 95% CI: 1.48–2.55), demonstrating progressive MACE risk across metabolic-inflammatory phenotypes (log-rank P<0.001). The shaded areas around the survival curves represent 95% confidence intervals.

Causal mediation analysis

Researchers carried out a mediation analysis in light of the possibility that coronary artery lesions might govern the biomarker–MACE relationship (Fig. 3) Regarding the HbA1c/HDL-c ratio, mediation analysis revealed that both the SYNTAX score and the rSS partially mediated its effect on MACE. The HbA1c/HDL-c ratio showed a significant effect on MACE. The proportion mediated by the SYNTAX score was 16.8% (95% CI: 1.95–38.0%), whereas the rSS mediated a greater proportion (25.2%, 95% CI: 12.0–52.0%), indicating that while both measures of coronary lesion complexity had a role in mediating the effect of the ratio on MACE, the HbA1c/HDL-c ratio’s direct effect persisted.

Sensitivity analyses demonstrated that the estimated mediation effects were robust to moderate levels of unmeasured confounding (ρ ranging from approximately − 0.3 to + 0.3) in all tested pathways. The results remained statistically significant across a broad range of ρ values, particularly for models involving the SYNTAX score and hsCAR, suggesting strong internal validity of the identified indirect effects. (Supplementary Fig. S4-5)

Subgroup analysis

Subgroup analysis (Fig. 4) revealed that both an elevated HbA1c/HDL-c ratio (≥ 6.61) and hsCAR (≥ 0.18) were independent risk factors with differential impacts across patient populations. Risk stratification was most pronounced in nondiabetic patients, normotensive individuals, and patients having raised SYNTAX scores. The prognostic significance of these biomarkers varied according to antiplatelet and lipid-lowering treatment regimens, indicating therapy-specific risk modifications.

Fig. 4.

Subgroup test of the link of MACE risk with combined biomarker phenotypes. Forest plot showing HR and 95% CIs for MACEs in prespecified subgroups. Patients were stratified into four phenotypes based on HbA1c/HDL-c and hsCAR levels, with the dual-negative group (HbA1c/HDL-c < 6.61 and hsCAR < 0.18) used as the reference. Each panel compares one biomarker group with the reference group across all subgroups. P for trend tests the significance across the four biomarker groups; P for interaction (P for int) assesses the heterogeneity of biomarker-associated risk across subgroups.

Notably, elevated hsCAR (≥ 0.18) consistently predicted adverse outcomes regardless of SGLT2i therapy (P < 0.001); however, the degree of risk was considerably greater among nonusers (odds ratio [OR] = 2.61) than among users (OR = 1.62), suggesting that SGLT2i may attenuate inflammation-mediated cardiovascular events.

ROC analysis and reclassification metrics

To gauge the incremental prognostic value of biomarker integration, ROC analysis was conducted to compare the combined HbA1c/HDL-c ratio + hsCAR model against individual biomarkers and the traditional GRACE score. The integrated model demonstrated superior discriminatory capacity (AUC: 0.63, 95% CI: 0.60–0.66; P < 0.001) (Fig. 5), significantly outperforming the GRACE risk scores (AUC: 0.56, 95% CI: 0.53–0.58; P < 0.001) alongside individual biomarkers when assessed separately. Notably, the combination yielded a significant integrated discrimination improvement (IDI: 0.02, 95% CI: 0.01–0.03; P < 0.01) (Table 3), reinforcing multimarker strategies’ clinical relevance in cardiovascular risk assessment for patients with STEMI following PCI. Similarly, there is a consistent result compared with that of the modified GRACE score (Supplementary Fig. S6 and Table S3).

Fig. 5.

Discriminatory capacity of individual and combined biomarker models for MACE prediction. Receiver operating characteristic curves comparing the prognostic value of the hsCAR alone (AUC 0.60, 95% CI: 0.57–0.63), GRACE risk score (AUC 0.56, 95% CI: 0.53–0.58), HbA1c/HDL-c ratio alone (AUC 0.58, 95% CI: 0.55–0.61), the combined HbA1c/HDL-c+ hsCAR model (AUC 0.63, 95% CI: 0.60–0.66; P<0.001 for comparison with all other models). The integrated model demonstrated significant improvement in risk reclassification (IDI: 0.02, 95% CI: 0.01–0.03; P<0.01)

Table 3.

Reclassification metrics analysis. Ninety-five percent confidence intervals show the area under the curve (AUC). Researchers computed the Integrated Discrimination Improvement (IDI) as well as the Net Reclassification Improvement (NRI) to find the incremental prognostic value of the hsCAR and HbA1c/HDL-c ratio beyond the baseline GRACE risk scores. The GRACE score served as the reference model. Researchers carried out a receiver operating characteristic curve test using DeLong's method for AUC comparisons. NRI was calculated using a risk threshold of 10% for event probability. Bootstrap resampling (1000 iterations) was used to calculate confidence intervals for NRI and IDI. *P < 0.05, **P < 0.01, ***P < 0.001 vs. GRACE risk score model

Model	AUC	P value	NRI	P value	IDI	P value
GRACE score	0.55 (0.53, 0.58)	Ref.	Ref.	Ref.	Ref.	Ref.
HbA1c/HDL-c	0.57 (0.54, 0.60)	0.57	0.20 (0.08, 0.31)**	< 0.01	0.01 (0.002, 0.02)*	< 0.05
hsCAR	0.60 (0.57, 0.63)*	< 0.05	0.08 (−0.04, 0.20)	0.22	−0.003 (−0.01, 0.006)	0.52
HbA1c/HDL-c + hsCAR	0.63 (0.60, 0.66)***	< 0.001	0.09 (−0.02, 0.21)	0.12	0.02 (0.01, 0.03)**	< 0.01

GRACE score Global Registry of Acute Coronary Events (GRACE) score, NRI Net Reclassification Improvement, IDI Integrated Discrimination Improvement

* denotes P < 0.05

** denotes P < 0.01, and

*** denotes P < 0.001

Prognostic value of the TyG index

Multivariable RCS test showed that the TyG index was nonlinearly related to MACE risk, with a threshold of 7.2, above which elevated TyG was significantly related to raised MACE risk (HR 1.58, 95% CI: 1.27–1.97; P < 0.001) (Supplementary Fig. S7). Additionally, combining the TyG index with hsCAR yielded a predictive model with a 0.64 AUC (95% CI: 0.61–0.67), similar to a 0.63 AUC (95% CI: 0.60–0.66) for the HbA1c/HDL-c ratio combined with hsCAR (Supplementary Fig. S8).

Sensitivity analysis of select bias

At the starting point, features of patients who completed follow-up, as well as those who were lost to follow-up, are shown in Supplementary Table S4. The distribution of sexes and ages was similar, but the group lost-to-follow-up had a lower rSS and SYNTAX score, a higher BMI, and fewer comorbidities (such as diabetes and hypertension). Moreover, these patients were less likely to get key drugs, such as ACEIs/ARBs/ARNIs, PCSK9is, and SGLT2is. Then, the team conducted an inverse probability weighting sensitivity analysis (Supplementary Table S5). The HbA1c/HDL-c ratio, hsCAR, and their combination’s weighted effects on MACE were consistent with the primary analysis.

Discussion

Here, researchers offer fresh perspectives on the predictive value of integrating inflammatory as well as metabolic biomarkers for STEMI risk stratification after PCI. By simultaneously evaluating the HbA1c/HDL-c ratio and hsCAR, a synergistic approach was demonstrated to significantly enhance the prediction of MACE. The identification of clinically relevant thresholds (HbA1c/HDL-c ratio ≥ 6.61; hsCAR ≥ 0.18) further enables practical implementation of these findings in clinical settings. This thorough examination, which takes causal mediation into account, shows how intricately systemic inflammation, metabolic dysfunction, cardiovascular outcomes, and coronary lesion complexity interact, offering new perspectives on pathophysiological mechanisms and potential therapeutic targets (Fig. 6).

Fig. 6.

Central illustration. Synergistic biomarker approach for enhanced MACE prediction after PCI in patients with STEMI: Combined HbA1c/HDL-c ratio and hsCAR risk stratification.

Integration of metabolic and inflammatory biomarkers

The HbA1c/HDL-c ratio reflects both long-term glycemic control and dyslipidemia. Since its introduction by Hu et al. in 2021 [7], it has shown prognostic value in conditions such as carotid atherosclerosis and stroke. This study extends its relevance to patients with STEMI undergoing PCI, a previously unstudied high-risk group.

Complementing this metabolic marker, hsCAR combines hsCRP and albumin to capture systemic inflammation. It has been validated in various cardiovascular settings, including chronic total occlusion and post-CABG atrial fibrillation [28]– [29]. In STEMI cases, the hsCAR outperforms individual inflammatory markers [13, 30], and the findings further support that it strongly indicates complications after PCI.

The significant incremental prognostic value achieved by combining these biomarkers represents a key finding of the study. Compared with either the marker alone or the established GRACE risk score, the integrated model demonstrated superior discriminatory capacity (AUC: 0.63, 95% CI: 0.60–0.66; P < 0.001). This enhanced performance suggests that simultaneously capturing metabolic dysfunction and inflammatory activation offers a more all-inclusive information of the basic pathophysiology than conventional risk stratification approaches. Although the AUC of 0.63 for the combined biomarker model reflects moderate discrimination, this efficacy exceeds that of both the GRACE score (AUC: 0.56) and the individual markers. Moreover, in high-risk STEMI cases, even modest improvements in risk stratification based on easily obtainable biomarkers can inform early clinical decision making. The significant improvement in integrated discrimination (IDI: 0.02, 95% CI: 0.01–0.03; P < 0.01) further validates the clinical relevance of this multimarker strategy.

The homeostasis model assessment of insulin resistance, along with the hyperinsulinemia-euglycemic clamp, shows strong link with the TyG index, which suggests insulin resistance. Researchers calculated it as ln[fasting triglycerides (mg/dL) × fasting plasma glucose (mg/dL)/2] [31]. The TyG index’s relationship with cardiovascular events has been well established [32]. Similar to the TyG index, the HbA1c/HDL-c ratio reflects long-term glycemic control and HDL-mediated cholesterol transport. These findings indicate that both markers exhibit comparable predictive capabilities for MACE.

Mechanistic insights from causal mediation analysis

The causal mediation analysis revealed important insights into the pathways linking these biomarkers to cardiovascular outcomes. Both biomarkers demonstrated significant direct effects on MACE, with coronary lesion complexity (SYNTAX score and rSS) serving as partial mediators. These findings suggest a dual impact of both biomarkers on cardiovascular risk: first, through direct pathophysiological mechanisms, and second, through their influence on coronary lesion burden. The stronger mediation of HbA1c/HDL-c’s impact through the rSS suggests that metabolic dysfunction may particularly affect incomplete revascularization outcomes, whereas inflammation (as reflected by hsCAR) may more strongly influence baseline coronary complexity. Importantly, the effects of the majority of both biomarkers remained direct, indicating that additional pathways beyond coronary anatomy contribute to adverse outcomes.

Pathophysiological framework

The combined prognostic value of the hsCAR alongside the HbA1c/HDL-c ratio likely reflects the interplay of metabolic dysfunction with systemic inflammation in cardiovascular disease. Elevated HbA1c indicates chronic hyperglycemia, which promotes oxidative stress, endothelial dysfunction, and glycation [33], whereas low HDL-c impairs reverse cholesterol transport and antioxidant defenses [34], together accelerating atherogenesis and thrombosis [35]– [36]. The hsCAR captures the inflammatory burden by combining hsCRP and albumin. High hsCRP reflects proinflammatory activity, and low albumin indicates diminished anti-inflammatory and nutritional reserves. Inflammation enhances endothelial activation, plaque vulnerability, platelet reactivity, and thrombogenicity, thereby increasing post-PCI risk [37–40].

This bidirectional relationship forms a vicious cycle: hyperglycemia drives inflammation via advanced glycation end-products (AGE-RAGE) signaling and oxidative stress, whereas inflammation worsens metabolic dysfunction through insulin resistance and lipid dysregulation [41]– [42]. By capturing both processes, the combined biomarker model offers a more integrated assessment of residual risk than either marker alone.

Clinical implications

The HbA1c/HDL-c ratio and hsCAR capture complementary pathophysiological pathways—namely, metabolic dysfunction and systemic inflammation—that are not fully addressed by traditional risk scores. Improving existing risk stratification by integrating these markers is clinically meaningful. A key strength of this approach lies in its practicality: HbA1c, hsCRP, HDL-c, and albumin are routinely measured in standard clinical settings, requiring no additional laboratory infrastructure. Therefore, the derived ratios can be readily incorporated into electronic health record (EHR) systems as automated calculation tools, enabling real-time identification of high-risk individuals.

This strategy aligns well with the principles of stratified medicine, where low-cost, easily accessible biomarkers can support tailored decision-making in acute coronary syndrome care pathways [43]. Notably, such integration could help prioritize early or complete revascularization in STEMI cases, a critical determinant of prognosis [44]. Furthermore, given that microvascular obstruction (MVO) contributes to complications even after successful epicardial reperfusion, the inflammatory and metabolic milieu reflected by these ratios may offer insight into individual susceptibility to MVO [45].

In addition to predicting MACEs, assessing bleeding risk is critical for comprehensive post-PCI management, as bleeding complications significantly impact patient outcomes [46–49]. The PRECISE-HBR score combines the Academic Research Consortium for High Bleeding Risk (ARC-HBR) criteria with the Predicting Bleeding Complications in Patients Undergoing Stent Implantation and Subsequent Dual Antiplatelet Therapy (PRECISE-DAPT) score. It is one example of a recent development in risk stratification that offers useful tools to find people susceptible to bleeding complications after PCI [50]. Integrating these bleeding risk models with the proposed metabolic-inflammatory biomarker approach (HbA1c/HDL-c ratio and hsCAR) could further enhance personalized treatment strategies, balancing ischemic and bleeding risks to optimize the duration and intensity of dual antiplatelet therapy.

Strengths and limitations

This study has more than a few strengths, namely its substantial sample size, comprehensive clinical characterization, systematic evaluation of coronary anatomy, and formal causal mediation analysis. The identification of specific biomarker thresholds enhances clinical applicability, whereas extensive adjustment for potential confounders strengthens the validity of the findings.

This study is limited in a number of ways. First, the team might have caused selection bias because of the single-center, retrospective design, which would have limited how broadly the findings could be applied. The multivariable adjustments and large, well-defined cohort did not rule out the possibility. Second, all biomarkers were measured at a single time point during the acute STEMI phase, which may not fully reflect the patients’ chronic metabolic or inflammatory status. Acute-phase physiological stress and early treatment may influence biomarker levels, potentially introducing variability and limiting the interpretation of causal relationships [51]– [52]. Third, other emerging biomarkers (e.g., N-terminal pro B-type natriuretic peptide, interleukin-6, and troponin dynamics) and detailed imaging parameters (such as infarct size or microvascular obstruction on cardiac magnetic resonance imaging) were not available in this dataset. As such, researchers could not explore their additive or mediating roles in conjunction with the hsCAR as well as HbA1c/HDL-c ratio.

Conclusions

When researchers combined the high-sensitivity C-reactive protein-to-albumin ratio and the HbA1c/HDL-c ratio, risk stratification in STEMI greatly improved following PCI by capturing complementary metabolic and inflammatory pathways, providing actionable biomarker thresholds that enable personalized therapeutic strategies and precision medicine applications in contemporary cardiovascular practice. Future prospective validation studies in diverse populations and randomized trials evaluating biomarker-guided therapeutic interventions are essential to confirm their clinical utility and establish whether enhanced risk stratification translates into improved patient outcomes.

Abbreviations

MACE

Major adverse cardiovascular events

CAD

Coronary artery disease

ACS

Acute coronary syndrome

CVD

Cardiovascular disease

STEMI

ST-elevation myocardial infarction

BMI

Body mass index

DBP

Diastolic blood pressure

BNP

B-type natriuretic peptide

HbA1c

Glycated hemoglobin A1c

FBG

Fasting blood glucose

HDL-C

High-density lipoprotein cholesterol

hsCRP

High-sensitivity C-reactive protein

LVEF

Left ventricular ejection fraction

LDL-C

Low-density lipoprotein cholesterol

SBP

Systolic blood pressure

TG

Triglyceride

Scr

Serum creatinine

TC

Total cholesterol

TNT

Troponin T

SYNTAX score

SYNergy between PCI with TAXUS and the cardiac surgery score

rSS

Residual SYNTAX score

PCI

Percutaneous coronary intervention

ARB

Angiotensin receptor blocker

ACEI

Angiotensin-converting enzyme inhibitor

GLP1-RA

Glucagon-like peptide-1 receptor agonist

ARNI

Angiotensin receptor and neprilysin inhibit

PCSK9i

Proprotein convertase subtilisin/kexin type 9 inhibitor

SGLT2i

Sodium-glucose cotransporter 2 inhibitor

MRA

Mineralocorticoid receptor antagonist

ROC

Receiver operating characteristic

CI

Confidence interval

RCS

Restricted cubic splines

HR

Hazard ratio

OR

Odds ratio

Authors’ contributions

SX and QY: Ideation, methodology, finance acquisition, and project management. JH: Methodology, gathering data, presentation, statistical tests, writing—creation of the first draft. JZ: Gathering data, formal test, presentation, and statistical validation. ZX and LL: Formal analysis, statistical validation, data interpretation. MC, YL, XY, SD, QW, and JC: Patient enrollment, data collection, clinical assessment, and follow-up coordination. SX and QY contributed equally as corresponding authors. JH and JZ contributed equally as first authors. The final draft of the paper has been read, edited, and cleared by all authors, who also pledge to take responsibility for every part of the work.

Funding

The Suzhou Industrial Park Xinxin Cardiovascular Health Foundation (Grant No. 2020-CAA-ACCESS-151) offered us funds. Study layout, gathering and interpreting data, manuscript planning, and the choice to submit the work for publishing were all outside the funding body’s purview.

Data availability

Due to institutional norms and privacy laws, the information sets created and examined during this study are not accessible to everyone. However, they can be obtained from the corresponding author upon a fair inquiry, provided that the institutional ethics committee has given its consent.

Declarations

Ethics approval and consent to participate

Tianjin Medical University General Hospital’s Ethics Committee gave its clearance for this study, which was carried out in compliance with the Declaration of Helsinki (approval number: IRB2023-YX-301-01/2023). The institutional review board exempted the need for each person’s written informed permission because this research was retrospective and used deidentified medical information.

Consent for publication

Not relevant because there is no personally identifying information in the paper.

Competing interests

The authors declare no competing interests.