Prognostic Utility of the Cytokine Profile for Predicting Major Adverse Cardiovascular Events in Acute Myocardial Infarction Following Percutaneous Coronary Intervention

Abstract

Objective

This study aimed to assess the clinical utility and prognostic value of a panel of serum cytokines for predicting major adverse cardiovascular events (MACE) in patients with acute myocardial infarction (AMI) after percutaneous coronary intervention (PCI).

Methods

We enrolled 232 AMI patients who underwent PCI. Based on the occurrence of MACE during follow-up, patients were categorized into a poor prognosis group (n=127) and a good prognosis group (n=105). Serum levels of interleukin IL-1β, IL-6, IL-8, IL-10, and tumor necrosis factor-alpha (TNF-α) were measured. Spearman correlation analysis was employed to evaluate the relationships between cytokine levels and MACE risk. Nine distinct machine learning models, incorporating all identified independent risk factors for MACE, were developed and evaluated.

Results

Levels of high-sensitivity C-reactive protein (hs-CRP), IL-6, IL-8, and TNF-α were significantly elevated in patients who developed MACE compared to those who did not. IL-6 emerged as a relatively strong predictor of MACE, with an area under the curve (AUC) of 0.82 (95% CI: 0.77–0.87). The IL-6/IL-10 ratio also demonstrated predictive value for MACE (AUC=0.74, 95% CI: 0.69–0.80). Furthermore, significant differences were observed between initial presentation and readmission levels for the IL-8/IL-10 and TNF-α/IL-10 ratios in patients readmitted for AMI following MACE. Among the nine machine learning models constructed, the XGBoost and Logistic Regression models, which incorporated all independent risk factors, demonstrated robust performance, both AUC values exceeded 0.8.

Conclusion

AMI patients with elevated serum cytokine levels exhibited a lower 1-year survival rate. Serum cytokine profiles show significant predictive value for MACE following PCI in AMI patients.

Introduction

Acute myocardial infarction (AMI) represents a critical cardiovascular emergency, with a rising global incidence that poses a significant threat to public health.^1,2 Percutaneous coronary intervention (PCI) is a primary reperfusion strategy for AMI, effectively relieving coronary stenosis and restoring myocardial perfusion.³ Despite its benefits, a considerable number of patients experience major adverse cardiovascular events (MACE) post-procedure, which are associated with substantial morbidity and mortality.⁴ Consequently, the identification of reliable predictors for MACE is crucial for risk stratification and improving clinical outcomes.

Given the central role of inflammation in the pathogenesis of atherosclerosis and the subsequent healing processes post-AMI,⁵ MACE has been established as a common composite endpoint in cardiovascular trials and is endorsed by regulatory agencies such as the U.S. Food and Drug Administration (FDA) and the European Medicines Agency (EMA) for evaluating the cardiovascular safety of new therapies.⁶ Research has increasingly focused on systemic inflammatory indices. For instance, the systemic immune-inflammation index (SII), which integrates lymphocyte, neutrophil, and platelet counts, has been linked to adverse outcomes in patients with ischemic heart disease and ST-elevation myocardial infarction (STEMI).^7,8 Similarly, the ratio of high-sensitivity C-reactive protein (hs-CRP) to albumin has demonstrated prognostic value for STEMI patients undergoing primary PCI.^9,10

Cytokines, as key mediators of inflammation, offer profound insights into post-AMI pathophysiology. Elevated levels of cytokines like interleukin IL-6 and tumor necrosis factor-alpha (TNF-α) have been associated with adverse cardiac remodeling and poor prognosis.^11,12 Notably, therapeutic inhibition of these inflammatory pathways is being investigated as a potential disease-modifying strategy following myocardial infarction.¹³ While previous studies have often examined single cytokines, the inflammatory response to AMI is highly complex and multifaceted.⁹ Therefore, a simultaneous analysis of multiple cytokines may provide a more comprehensive prognostic assessment than any single marker alone.

This study aims to evaluate the clinical utility and prognostic significance of a panel of serum cytokines for predicting MACE in AMI patients following PCI. By retrospectively analyzing patient data and employing advanced statistical models, we seek to identify key inflammatory risk factors to better inform clinical management and improve long-term prognosis.

Materials and Methods

Study Population

A retrospective analysis was conducted on 232 patients who underwent percutaneous coronary intervention for AMI at Wuhan Asia Heart Hospital between January 2023 and December 2024. The cohort comprised 176 males and 56 females. Patients were followed during hospitalization and for one year after discharge via outpatient visits or telephone contact. The follow-up endpoint was defined as either the completion of the one-year follow-up period or the occurrence of MACE, including cardiac death, all-cause mortality, recurrent myocardial infarction, malignant arrhythmia, recurrent angina pectoris, or acute heart failure.The study was conducted in accordance with the principles of the Helsinki Declaration and received approval from the Ethics Committee of Wuhan Asia Heart Hospital (2025-B047). Owing to the retrospective design, the requirement for written informed consent was waived, all patient data included in this research comply with confidentiality requirements.

Data Collection

Clinical data were extracted from the Electronic Medical Record System of Wuhan Asia Heart Hospital. The collected information included patient demographics, comorbidities, clinical presentation, laboratory findings, treatment strategies, and outcomes. Laboratory assessments covered routine blood tests, blood biochemistry, inflammatory markers, cardiac biomarkers, coagulation profiles, and cytokine levels. Estimated glomerular filtration rate (eGFR) was calculated using the CKD-EPI 2009 equation for all study participants.

Inclusion and Exclusion Criteria

Inclusion Criteria: (1) Diagnosis of AMI according to the *2019 Guidelines for the Diagnosis and Treatment of Acute ST-Segment Elevation Myocardial Infarction*¹⁴ and the *2024 Guidelines for the Diagnosis and Treatment of Non-ST-Segment Elevation Acute Coronary Syndrome*;¹⁵ (2) Presence of ischemic symptoms with a symptom-to-procedure time of less than 12 hours; (3) Age over 18 years; (4) Availability of complete clinical data, including inflammatory markers such as complete blood count results.

Exclusion Criteria: (1) History of previous myocardial infarction, prior stent implantation, or coronary artery bypass grafting; (2) Receipt of thrombolytic therapy before PCI; (3) Comorbid cardiac conditions, including congenital heart disease, cardiomyopathy, myocarditis, rheumatic heart disease, infective endocarditis, valvular heart disease, or pericardial disease; (4) Concomitant cerebrovascular disease (eg., cerebral infarction, cerebral hemorrhage), thyroid disorders, active infection, autoimmune diseases, malignancy, tuberculosis, or hematological diseases; (5) Severe hepatic or renal dysfunction, or recent use of medications known to affect liver function; (6) Incomplete clinical data.

Statistical Analysis

Statistical analyses were performed using GraphPad Prism version 8 (San Diego, CA, USA) and SPSS version 26.0 (Chicago, IL, USA). Continuous variables with a normal distribution are presented as mean ± standard deviation (SD) and were compared using the Student’s t-test. Non-normally distributed continuous data are presented as median (interquartile range, IQR) and were compared using the Mann–Whitney U-test, a non-parametric test. Comparisons of paired data were conducted using the paired t-test. Categorical variables are expressed as numbers (percentages) and were compared using the Chi-square test.

Spearman correlation analysis was used to examine the relationships between cytokine levels and MACE risk. Univariate and multivariable Cox regression analyses were performed to identify independent factors associated with short-term MACE occurrence in AMI patients.

The predictive performance of cytokine levels for short-term MACE was evaluated using receiver operating characteristic (ROC) curve analysis. The area under the curve (AUC) was interpreted as follows: <0.7, poor; 0.7–0.8, moderate; 0.8–0.9, good; and ≥0.9, excellent discriminatory power.

Machine learning models for MACE risk prediction were developed using Python (xgboost=2.0.1; lightgbm=3.2.1; scikit-learn=1.1.3). A two-sided p-value < 0.05 was considered statistically significant for all analyses.

Results

Baseline Patient Characteristics

A total of 232 acute myocardial infarction patients treated with percutaneous coronary intervention (PCI) were included in this study. Based on the occurrence of MACE during follow-up, patients were categorized into a MACE group (n=127) and a non-MACE group (n=105), the demographic and clinical characteristics of the two patient groups upon admission are presented in Table 1. No significant differences (P > 0.05) were observed between the two groups regarding age, sex, height, weight, or body mass index (BMI). However, significant differences were identified in heart rate, AMI type, and Killip classification.

Table 1.

Demographic and Clinical Characteristics on Admission of Patients with Acute Myocardial Infarction After PCI

Parameters	Total (n = 232)	non-MACE (n = 105)	MACE (n = 127)	Statistic	p
Age, mean (±SD), year	63±11	61±11	64±12	−1.92	0.056
Sex, n (%)				1.06	0.303
Female	56(24.1)	22(21.0)	34(26.8)
Male	176(75.9)	83(79.1)	93(73.2)
Underlying disease, n (%)
Hypertension	156(67.2)	67(63.8)	89(70.1)	1.03	0.311
Diabetes mellitus	80(34.5)	33(31.4)	47(37.0)	0.79	0.374
Hyperlipidemia	80(34.5)	36(34.3)	44(34.7)	0.00	0.954
Drinking history, n (%)	20(8.6)	10(9.5)	10(7.8)	0.20	0.656
Smoking history, n (%)	106(45.7)	45(42.9)	61(48.0)	0.62	0.431
Heart rate, median [IQR], beats/minute	78[68,88]	76[66,85]	80[71,90]	−2.60	0.009
Height, median[IQR], cm	168[162,172]	169[162,172]	168[161,172]	0.93	0.351
Weight, mean (±SD), kg	69.0±12.1	69.8±11.4	68.3±12.6	0.92	0.361
BMI, median [IQR], kg/m2	24.5[22.0,26.7]	24.5[22.6,26.8]	24.2[21.6,26.5]	0.83	0.410
AMI Type, n (%)				13.549	<0.001
STEMI	117(50.4)	39(37.1)	78(61.4)
NSTEMI	115(49.6)	66(62.9)	49(38.6)
Killip class, n (%)				26.58	<0.001
1	170(73.3)	93(88.6)	77(60.6)
2	27(11.6)	8(7.6)	19(15.0)
3	11(4.7)	3(2.9)	8(6.3)
4	24(10.3)	1(1.0)	23(18.1)
Outcome, N (%)				9.553	0.002
Survival	221(95.3)	105(100.0)	116(91.3)
Deceased	11(4.7)	0(0.00)	11(8.7)

Note: Bold pvalues are statistically significant.

Abbreviations: MACE, Major Adverse Cardiovascular Events; BMI, Body Mass Index; AMI, Acute Myocardial Infarction.

The distribution of AMI types differed significantly between the groups (χ² = 13.549, p < 0.001). While the overall cohort was nearly evenly split between ST-segment elevation myocardial infarction (STEMI, 50.4%) and non-ST-segment elevation myocardial infarction (NSTEMI, 49.6%), their distribution was markedly skewed. The MACE group was predominantly composed of STEMI patients (61.4%), whereas the non-MACE group had a significantly higher proportion of NSTEMI patients (62.9%).

Baseline clinical severity, as assessed by the Killip classification, also revealed pronounced disparities. The majority of the overall cohort (73.3%) was in Killip class I. The non-MACE group comprised a predominantly low-risk population, with 88.6% of patients in Killip class I and only 1.0% in class IV. In contrast, the MACE group represented a higher-risk population. Although Killip class I remained the most frequent category, its prevalence (60.6%) was substantially lower. This group exhibited a higher burden of heart failure, with greater proportions of patients in Killip classes II (15.0% vs. 7.6%), III (6.3% vs. 2.9%), and IV (18.1% vs. 1.0%), indicating significantly greater illness severity at presentation.

A statistically significant disparity in mortality outcomes was also observed (χ² = 9.553, p = 0.002). The overall cohort mortality rate was 4.7%. The non-MACE group demonstrated 100% survival (0.00% mortality), whereas the MACE group experienced a significantly higher mortality rate of 8.7%.

Analysis of Inflammatory Markers in AMI Patients After PCI

Serum levels of inflammatory markers were compared at initial admission between patients who subsequently developed MACE and those who did not (Figure 1). Levels of hs-CRP, IL-6, IL-8, and TNF-α were significantly elevated in the MACE group compared to the non-MACE group (P < 0.001; Figure 1A). In contrast, no significant inter-group differences were observed in the levels of IL-1β and IL-10 (P > 0.05). Analysis of cytokine ratios revealed that the IL-6/IL-10 ratio was also significantly higher in the MACE group (P < 0.001; Figure 1B).

Figure 1.

Analysis of serum inflammatory marker levels at admission in patients with and without MACE. (A) Comparison of cytokine levels between the two groups; (B) Comparison of cytokine ratios between the two groups. ****P < 0.0001.

Abbreviation: ns, not significant.

These findings indicate that patients who developed MACE presented with a more pronounced systemic inflammatory state at admission compared to those without such events. The elevated levels of specific cytokines and ratios upon hospitalization suggest their potential utility as early biomarkers for predicting long-term cardiovascular risk after PCI.

We further performed ROC curve analysis to evaluate the predictive capacity of the differentially expressed inflammatory markers for MACE (Figure 2A). IL-6 demonstrated considerable predictive ability for post-PCI MACE, with an AUC of 0.82 (95% CI: 0.77–0.87). The IL-6/IL-10 ratio also showed significant predictive value, with an AUC of 0.74 (95% CI: 0.69–0.80).

Figure 2.

ROC curve analysis of the predictive ability of inflammatory markers for MACE. (A) ROC curve results of inflammatory markers. (B) ROC curve results of previously reported other established inflammatory and ratio indices.

Subsequently, the predictive performance of IL-6 was compared with that of other established inflammatory and ratio indices, including the systemic immune-inflammation index (SII), neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), lymphocyte-to-monocyte ratio (LMR), hemoglobin, albumin, lymphocyte and platelet (HALP) score, and C-reactive protein-to-albumin ratio (CAR) (Figure 2B). The analysis revealed that IL-6 exhibited superior discriminatory performance compared to all these benchmarks.

In summary, among all indicators assessed, IL-6 emerged as the most robust single predictor for MACE risk, demonstrating greater predictive accuracy than other commonly reported inflammatory indices.

Association of Baseline Cytokine Levels with One-Year MACE and Mortality

Spearman correlation analysis was conducted to evaluate the associations between cytokine levels at admission and the one-year risk of MACE in AMI patients. As detailed in Table 2, levels of IL-6, IL-8, and TNF-α, as well as the IL-6/IL-10 and IL-8/IL-10 ratios, demonstrated significant positive correlations with MACE risk.

Table 2.

Correlation Analysis Between Cytokine Levels at Admission and the Risk of MACE within One Year in Patients with AMI

Parameters	IL-1β	TNF-α	IL-10	IL-6	IL-8	IL-1β/IL-10	IL-6/IL-10	IL-8/IL-10	TNF-α/IL-10
r	0.024	0.209	0.048	0.258	0.206	-0.010	0.230	0.132	0.122
P	0.714	0.001	0.472	<0.001	0.002	0.878	<0.001	0.044	0.064

Note: Bold p values are statistically significant.

Furthermore, we investigated the relationship between cytokine levels and mortality among patients who experienced a MACE. A comparison of the survival and non-survival subgroups within the MACE cohort revealed that non-survivors had significantly elevated levels of IL-8 compared to survivors (P = 0.020), while no statistically significant differences were observed in the remaining cytokines (P>0.05),as detailed in Supplementary Table 1.

Dynamic Changes in Cytokine Levels Following Post-PCI MACE

To investigate the inflammatory trajectory after adverse events, we compared cytokine levels at initial admission and upon readmission in AMI patients who experienced a post-PCI MACE (Figure 3). At readmission, TNF-α levels were significantly elevated compared to baseline (17.09 ± 12.83 pg/mL vs. 13.08 ± 7.27 pg/mL, P = 0.007). Furthermore, the IL-8/IL-10 and TNF-α/IL-10 ratios were also significantly higher at readmission than at initial presentation (P < 0.05).

Figure 3.

Changes in cytokine levels in AMI patients who developed post-PCI MACE between initial admission and readmission.The red lines represent the trajectories of cytokine levels in individual patients between initial admission and readmission. *P < 0.05, **P < 0.01.

Abbreviation: ns, not significant.

These findings indicate that in AMI patients who experience a post-PCI MACE, the systemic inflammatory state progresses adversely after the event. This progression is characterized by a further increase in pro-inflammatory drive, evidenced by elevated TNF-α, and a heightened imbalance between pro- and anti-inflammatory pathways, as reflected by the increased cytokine ratios.

Identification of Independent Risk Factors for MACE

To identify independent risk factors for MACE, we applied the LASSO regression model to screen significant variables from clinical features and laboratory parameters that differed at admission (Supplementary Table 2). At the optimal lambda (λ) value of 0.02 (minimum mean squared error), the model selected the following variables: Killip class, IL-8, TNF-α, IL-6/IL-10 ratio, LVEDD, LVEF, NT-proBNP, hs-cTnI, NEUT, EO, ESR, Alb, A/G, eGFR, Creatinine, UA, hs-CRP, and HDL-C (Figure 4A). Under the more stringent criterion (λ = 0.031, one standard error of the minimum), the model selected a more parsimonious set of variables: Killip class, TNF-α, IL-6/IL-10 ratio, LVEDD, LVEF, NT-proBNP, hs-cTnI, NEUT, EO, ESR, Alb, AG, eGFR, hs-CRP, and HDL-C.

Figure 4.

(A) The LASSO regression model was applied to screen for significant variables from the clinical features and laboratory parameters that differed at admission in patients who experienced MACE. Ten-fold cross-validation was used to determine the tuning parameter λ. The left panel shows the LASSO regression coefficient path plot, where each colored curve represents the regression coefficient of a feature. The right panel displays the Binomial Deviance for candidate features as a function of λ, with the two vertical lines indicating the optimally simplified model within the standard error range. The red curve represents the mean deviance, and the gray area represents the confidence interval of the deviance. (B) ROC curve results for different models in the training and validation sets.

Subsequently, univariate and multivariable Cox regression analyses were performed on the pooled variables identified by both LASSO criteria. Multivariable analysis confirmed that the Killip class, TNF-α, NT-proBNP, ESR, ALb, A/G and eGFR were independent risk factors for MACE in AMI patients after PCI (Table 3).

Table 3.

Univariate and Multivariable Cox Regression Analysis of Risk Factors for MACE in Patients After PCI for Acute Myocardial Infarction

Parameters	Univariate		Multivariable
	P-value	HR (95% CI)	P-value	HR (95% CI)
Killip class
2	0.005	2.077(1.254,3.443)	0.006	2.358(1.232,4.511)
3	0.060	2.012 (0.971,4.172)	0.595	1.356(0.588,3.130)
4	<0.001	4.123(2.569,6.618)	0.120	2.043(1.015,4.111)
TNF-α	0.001	1.024(1.014,1.034)	0.098	1.032(0.994,1.072)
IL-6/IL-10	0.008	1.018(1.007,1.028)	0.145	0.969(0.928,1.011)
LVEDD	<0.001	1.518(1.113,2.072)	0.870	1.033(0.698,1.530)
LVEF	<0.001	0.931(0.908,0.954)	0.108	0.969(0.932,1.007)
NT-proBNP	<0.001	1.000(1.000,1.000)	0.002	1.000(1.000,1.000)
Hs-cTnI	<0.001	1.002(1.001,1.003)	0.500	0.999(0.997,1.001)
NEUT	<0.001	1.125(1.078,1.175)	0.627	1.066(0.824,1.378)
ESR	<0.001	1.020(1.013,1.027)	0.006	1.015(1.004,1.026)
ALb	<0.001	0.898(0.873,0.924)	0.009	0.902(0.835,0.975)
A/G	<0.001	0.118(0.057,0.246)	0.022	0.319(0.120,0.849)
eGFR	0.030	0.984(0.977,0.990)	0.041	0.988(0.977,1.000)
hsCRP	0.005	1.010(1.007,1.013)	0.281	1.004(0.997,1.011)
HDL-C	0.001	0.365(0.180,0.739)	0.106	0.446(0.167,1.187)
EO	<0.001	0.010(0.001,0.161)	0.124	0.068(0.002,2.079)

Note: Bold p values are statistically significant.

Abbreviations: LVEDD, Left Ventricular End-Diastolic Diameter; LVEF, Left Ventricular Ejection Fraction; NT-proBNP, N-terminal pro-B-type Natriuretic Peptide; Hs-cTnI, High-sensitivity cardiac Troponin I; NEUT, Neutrophil Count; ESR, Erythrocyte Sedimentation Rate; hsCRP, High-sensitivity C-reactive Protein; ALb, Albumin; A/G, Albumin-to-Globulin Ratio; eGFR, estimated Glomerular Filtration Rate; HDL-c, High-density Lipoprotein; EO, Eosinophils Count; HR, hazard ratios.

Predictive Performance of Machine Learning Models for MACE

Nine distinct machine learning models were constructed, all incorporating the independent risk factors identified for MACE. Model performance was evaluated using a 15-fold cross-validation approach, with the ROC curves for the training and validation sets depicted in Figure 4B. In the validation set, both the XGBoost and Logistic Regression models demonstrated robust predictive ability,which achieved the highest AUC of 0.82. The comprehensive performance metrics for all nine models are detailed in Table 4. Performance of Machine Learning Models Built on independent Risk Factors for MACE in the Validation Set.

Table 4.

Analysis of Model Performance Metrics

Model	AUC (95% CI)	Cutoff (95% CI)	Accuracy (95% CI)	Sensitivity (95% CI)	Specificity (95% CI)	PPV (95% CI)	NPV (95% CI)	F1 Score (95% CI)	Kappa (95% CI)
XGBoost	0.82 (0.60–0.99)	0.49(0.46–0.51)	0.75(0.70–0.81)	0.78(0.71–0.85)	0.72(0.63–0.82)	0.79(0.73–0.85)	0.75(0.68–0.82)	0.77(0.72–0.82)	0.50(0.40–0.61)
Logistic	0.81 (0.59–0.98)	0.55(0.54–0.56)	0.74(0.69–0.79)	0.66(0.57–0.74)	0.84(0.77–0.90)	0.85(0.79–0.91)	0.69(0.63–0.77)	0.73(0.66–0.79)	0.49(0.39–0.58)
RandomForest	0.79 (0.60–0.99)	0.63(0.62–0.64)	0.69(0.64–0.74)	0.59(0.50–0.68)	0.81(0.73–0.89)	0.81(0.75–0.88)	0.64(0.57–0.70)	0.67(0.61–0.73)	0.39(0.30–0.49)
AdaBoost	0.79 (0.60–0.99)	0.49(0.49–0.50)	0.730(0.67–0.78)	0.7(0.61–0.79)	0.76(0.69–0.83)	0.79(0.73–0.85)	0.70(0.63–0.78)	0.73(0.67–0.79)	0.46(0.36–0.56)
DecisionTree	0.76 (0.52–0.97)	0.63(0.60–0.67)	0.70(0.65–0.74)	0.65(0.57–0.73)	0.75(0.65–0.86)	0.79(0.72–0.85)	0.66(0.59–0.73)	0.69(0.65–0.74)	0.39(0.31–0.48)
KNN	0.72(0.48–0.95)	0.97(0.90–1.03)	0.66(0.61–0.70)	0.49(0.39–0.59)	0.86(0.76–0.96)	0.86(0.78–0.94)	0.59(0.55–0.63)	0.59(0.51–0.69)	0.33(0.25–0.42)
LightGBM	0.67(0.39–0.93)	0.50(0.39–0.61)	0.66(0.61–0.71)	0.62(0.54–0.71)	0.71(0.63–0.78)	0.72(0.67–0.77)	0.630(0.55–0.71)	0.66(0.60–0.72)	0.32(0.22–0.42)
GBDT	0.67 (0.45–0.88)	0.94(0.92–0.95)	0.65(0.60–0.70)	0.49(0.40–0.57)	0.85(0.77–0.92)	0.83(0.75–0.91)	0.59(0.54–0.64)	0.59(0.53–0.66)	0.32(0.23–0.41)
SVM	0.65(0.36–0.91)	0.61(0.59–0.64)	0.63(0.60–0.70)	0.68(0.59–0.76)	0.56(0.44–0.68)	0.67(0.60–0.74)	0.59(0.49–0.69)	0.66(0.60–0.72)	0.24(0.10–0.38)

Abbreviations: PPV, Positive Predictive Value; NPV, Negative Predictive Value.

Discussion

Despite the widespread application of PCI for effectively relieving coronary stenosis, a substantial number of patients remain at risk for MACE post-procedure.^16,17 Inflammation is a key driver of these complications. Elevated levels of hs-CRP signify a pronounced inflammatory response within the coronary intima, which enhances plaque instability and predisposes to adverse cardiovascular events.^18,19 Cytokines, as central mediators of inflammation, have been strongly implicated in adverse cardiac remodeling and outcomes following AMI. The present study systematically investigated the inflammatory profiles and their dynamic changes in AMI patients undergoing PCI, integrating traditional clinical parameters, cytokine profiling, and machine learning approaches to elucidate the pathophysiological basis of post-infarction complications and improve risk stratification.

In alignment with previous studies, our study found that levels of hs-CRP, IL-6, IL-8, and TNF-α, as well as the IL-6/IL-10 ratio, were significantly elevated at admission in patients who subsequently developed MACE. This indicates that a more intense systemic inflammatory state is already present at baseline in these high-risk individuals. IL-10 plays a protective role in atherogenesis and plaque stability in animal models by limiting inflammatory injury and promoting tissue repair,^20–22 the elevated IL-6/IL-10 ratio in our MACE group likely reflects a state of relative IL-10 insufficiency, creating a pro-inflammatory milieu conducive to adverse events. This finding suggests that MACE risk is associated not only with the absolute magnitude of pro-inflammatory activation but also with an imbalance between pro- and anti-inflammatory mechanisms. The clinical literature on IL-10 in acute coronary syndromes (ACS) has yielded complex findings. Some studies report reduced IL-10 levels in ACS patients compared to stable counterparts,²³ and one study found that elevated baseline IL-10 predicted a reduced risk of death and MI at 6 months.²⁴ However, a larger subsequent study paradoxically found that elevated IL-10 independently predicted increased risk of death and MI at one year.²⁵ Our data suggest that the prognostic information carried by IL-10 may be context-dependent and best captured when considered in relation to pro-inflammatory cytokines, as the IL-6/IL-10 ratio proved more informative than IL-10 levels alone.

IL-8, a CXC chemokine, is crucial for recruiting and activating neutrophils and monocytes and may promote angiogenesis.^26,27 Prior studies have reported elevated IL-8 levels in patients with ACS compared to those with stable angina or healthy controls,²⁸ and high levels are associated with an increased risk of adverse outcomes.^26,29 A recent study confirmed that high IL-8 levels predict future MACE in MI patients, independent of other risk factors.³⁰ In our cohort, absolute IL-8 levels were significantly different between the MACE and non-MACE groups at baseline, both IL-8 and the IL-8/IL-10 ratio showed a positive correlation with MACE risk. More importantly, within the MACE group, non-survivors exhibited significantly higher IL-8 levels than survivors. This finding links IL-8 specifically to fatal outcomes. Our observation aligns with a large prospective study demonstrating that elevated IL-8 independently predicts long-term all-cause mortality in ACS patients.²⁶ Collectively, these data position IL-8 as a potentially valuable biomarker for identifying, among patients who have already sustained a MACE, those at highest imminent risk of mortality.This suggests that elevated IL-8 may not merely be a risk marker but could actively contribute to mortality risk among those who experience a MACE, likely by exacerbating the pro-/anti-inflammatory imbalance, as reflected in the elevated cytokine ratios at readmission.

Our findings corroborate the prognostic value of IL-6, demonstrating its superior predictive performance (AUC = 0.82) compared to other established inflammatory indices such as the systemic immune-inflammation index (SII) and neutrophil-to-lymphocyte ratio (NLR). While easily obtainable cellular indices like the NLR, platelet-to-lymphocyte ratio (PLR), and SII have demonstrated prognostic value in coronary artery disease and ACS.^31–33 In the inflammatory phase following ST-segment elevation myocardial infarction (STEMI), processes involving IL-6 and neutrophil extracellular trap (NET) formation are highly active, amplifying tissue damage.³⁴ Elevated IL-6 plasma levels are associated with early and late stent thrombosis in clopidogrel-treated patients, suggesting a role in thrombotic complications.³⁵ Our analysis found that IL-6 outperformed these established ratios in predictive capacity. This suggests that direct measurement of specific proximal cytokines may provide more refined insight into the pathophysiological inflammatory state than composite cellular indices, which are indirect proxies for systemic inflammation.

A novel aspect of our study is the longitudinal comparison of cytokine profiles in patients readmitted following a MACE. Serial analysis revealed a dynamic progression of inflammation: in patients readmitted post-MACE, TNF-α levels and the IL-8/IL-10 and TNF-α/IL-10 ratios were significantly higher than at initial presentation. This dynamic pattern provides insight into the evolving inflammatory landscape following adverse events. Experimental studies have documented a delayed surge in pro-inflammatory cytokines following myocardial ischemia-reperfusion. In a rat model of reperfused myocardial infarction, Moro and colleagues observed a marked increase in myocardial TNF-α, IL-1β, and IL-6 at day 8 post-infarction, while anti-inflammatory cytokines IL-4 and IL-10 remained unchanged.³⁶ This delayed inflammatory response coincided with the development of left ventricular dysfunction and remodeling. Our clinical observation of elevated TNF-α at readmission aligns with this experimental time course, suggesting that MACE events may trigger or coincide with a secondary wave of inflammation characterized by TNF-α predominance. The persistently elevated cytokine ratios further indicate that the pro-inflammatory/anti-inflammatory imbalance not only persists but may intensify following adverse events, potentially creating a permissive environment for recurrent ischemia or heart failure progression.

Our findings are further strengthened by the application of machine learning. The robust performance of our logistic regression model (AUC = 0.82), which incorporated the identified inflammatory and clinical risk factors, is comparable or superior to previous machine learning studies for MACE prediction in MI patients.^37,38 This underscores the effectiveness of machine learning in cardiovascular risk stratification and supports the integration of cytokine profiles into such models for enhanced predictive accuracy.

Limitations

Our study has several limitations. First, its single-center, retrospective, observational design with a modest sample size may introduce selection bias and limits generalizability; the findings require validation in larger, multicenter prospective cohorts. Second, follow-up was partially conducted via telephone, which is susceptible to patient recall and personnel interpretation biases. Third, we did not include coronary angiography data or detailed plaque characteristics, which could provide additional prognostic information. It is also important to note the heterogeneity in the types of AMI between the groups, with a higher proportion of STEMI patients in the MACE group and NSTEMI patients in the non-MACE group. As previous studies have reported that the type of myocardial infarction, as well as the nature and location of the underlying plaque, can affect inflammatory marker levels, the role of these cytokines as independent objective prognostic factors warrants further investigation in more homogenously designed prospective studies. Fourth, MACE represents a composite endpoint encompassing death, reinfarction, and heart failure; different event types may be associated with distinct inflammatory phenotypes, and sample size constraints prevented subgroup analyses by event type. Finally, the 1-year follow-up period is relatively short; longer-term studies are warranted to fully ascertain the predictive value of cytokines in AMI patients.

Conclusion

In conclusion, AMI patients who experience MACE exhibit elevated serum levels of pro-inflammatory cytokines and a disrupted pro-/anti-inflammatory balance, both at presentation and during disease progression. These patients also demonstrate a lower 1-year survival rate. Our findings, validated by machine learning models, indicate that serum cytokine profiles hold significant predictive value for MACE following PCI in AMI patients, offering a potential tool for improved risk stratification.

Funding Statement

The research was supported by the Noncommunicable Chronic Diseases-National Science and Technology Major Project (2023ZD0503903).

Abbreviations

MACE, major adverse cardiovascular events; AMI, acute myocardial infarction; PCI, percutaneous coronary intervention; hs-CRP, high-sensitivity C-reactive protein; AUC, area under the curve; SVM, support vector machine; EMA, European Medicines Agency; SII, systemic immune-inflammation index; STEMI, ST-elevation myocardial infarction; TNF-α, umor necrosis factor-alpha; SD, standard deviation; ROC, receiver operating characteristic; BMI, body mass index; NLR, neutrophil-to-lymphocyte ratio; PLR, platelet-to-lymphocyte ratio; LMR, lymphocyte-to-monocyte ratio; CAR, C-reactive protein-to-albumin ratio; NET, neutrophil extracellular trap; ACS, acute coronary syndrome.

Data Sharing Statement

All data supporting the findings of this study are available within the paper and its Supplementary Information.

Ethics Approval and Consent to Participate

The study was approved by the Ethics Committee of Wuhan Asia Heart Hospital (2025-B047).The research was in compliance with the Helsinki Declaration and the informed consent to participate was waived due to the retrospective design in the study.

Disclosure

Juan Yang and Bo Zhang are co-first authors for this study. The authors declare no conflicts of interest in this work.