This multicenter prospective cohort study enrolled consecutive STEMI patients undergoing pPCI in 4 centers (Beijing Anzhen Hospital, Beijing Luhe Hospital, Civil Aviation General Hospital, Beijing, and The First People's Hospital of Lianyungang) from August 2019 to September 2021. Patients who underwent cardiac magnetic resonance (CMR) at the index admission (3-7 days after pPCI) were selected. After the exclusion of patients with insufficient image quality for functional angiographic analysis, patients with both angio-IMR and CMR were included in the analysis. Ethics approval was obtained from the institutional review boards of all participating centers, with written informed consent acquired from all participants prior to enrollment.

Diagnosis of STEMI adhered to the 2018 fourth universal definition of myocardial infarction (MI).12,13 pPCI procedures were performed in accordance with standardized protocols, with interventional strategies, including stent selection, thrombectomy use, and administration of glycoprotein IIb/IIIa inhibitors, determined by the operator's clinical judgment based on angiographic findings and hemodynamic status.

All CMR imaging was performed on 3.0T clinical scanners (Philips Ingenia CX, Netherlands; GE MR750W, USA). A standardized imaging protocol included balanced steady-state free precession cine sequences, black blood fat-suppressed T2-weighted imaging, and late gadolinium enhancement. Full imaging parameters are detailed in the methods of the supplementary data. Two independent analysts (Q. Guo and Y. Guo), trained in advanced cardiac imaging (≥ 3 years of experience), conducted blinded quantitative assessments supervised by a board-certified cardiac radiologist (H. Wang). Comprehensive methodological details are available in the methods of the supplementary data.

Angio-IMR measurements were derived from angiographic projections meeting specific criteria: ≥ 2 orthogonal views separated by ≥ 30° without table movement, avoiding vessel overlap. Aortic pressure waveforms acquired post-pPCI and angiographic data were processed using FlashAngio IMR software (Rainmed Ltd, Suzhou, China) to generate 3-dimensional coronary artery reconstructions spanning from the inlet to the most distal position. Diastolic flow velocity Vdiastole was calculated via the Thrombolysis in Myocardial Infarction (TIMI) Frame Count method, defined as the ratio of contrast agent travel distance in 3-dimensional-reconstructed coronaries during the period of diastole. Maximal hyperemic flow velocity Vhyp was estimated as 2.1×Vdiastole14 A validated computational fluid dynamics model simulated steady-state laminar flow across stenotic vessels within 10 to 30seconds.15 Angio-fractional flow reserve values were computed using established algorithms,15 and angio-IMR was calculated via the formula:

where L=75, a dimensionless constant corresponding to 75mm downstream from the inlet to the distal position. Analyses of all 3 major vessels were performed by an independent core laboratory blinded to clinical and CMR data. The thresholds for CMD were defined as angio-IMR> 40 U in culprit vessels consistent with prior publications.16 Although angio-IMR values were obtained in both culprit and nonculprit vessels, only the measurements from the culprit artery were included in the statistical analyses. Full procedural protocols are detailed in the methods of the supplementary data.

MetS was defined using adult treatment panel III criteria.17 The exception was that body mass index (BMI) was used as a surrogate parameter for waist circumference which was not available at baseline. BMI is included as a diagnostic criterion for MetS within several recognized classification systems.18 Patients were classified as having MetS in the presence of 3 or more of the following components: a) BMI ≥ 28kg/m2, according to Chinese standard19,20; b) triglycerides ≥ 150mg/dL; c) high-density lipoproteins cholesterol <40mg/dL in men or <50mg/dL in women; d) blood pressure of at least 130/85mmHg or use of antihypertensive medication; beta-blockers and angiotensin-converting enzyme inhibitors or angiotensin receptor blockers were considered as antihypertensive therapy only if a diagnosis of hypertension was indicated by the investigator; e) fasting plasma glucose concentration ≥ 100mg/dL or use of diabetic medications.

Patients were prospectively followed up through November 2023, with follow-up evaluations scheduled at 1, 3, 6, and 12 months following STEMI, and every 6 months thereafter. Clinical data were systematically collected through structured outpatient visits, hospitalization records, or standardized telephone interviews. All reported events were rigorously verified against electronic medical records and independently reviewed by cardiologists blinded to CMR results to ensure unbiased adjudication.

The primary endpoint, major adverse cardiovascular and cerebrovascular events (MACCE), comprised a composite outcome including cardiac death, recurrent MI, stroke, ischemia-driven revascularization (repeat PCI or coronary artery bypass grafting), malignant ventricular arrhythmias (ventricular fibrillation, sustained ventricular tachycardia, or cardiac arrest), and hospitalization for heart failure. Secondary endpoints focused on individual components of MACCE, LV dysfunction and lack of LV ejection fraction (LVEF) recovery at follow-up. All definitions of outcomes strictly adhered to the Standardized Data Collection for Cardiovascular Trials Initiative criteria. Stroke was defined as an acute symptomatic episode of neurologic dysfunction, including ischemic stroke and hemorrhagic stroke. Reinfarction was diagnosed using the same criteria as for MI. Ischemia-driven revascularization encompassed any repeat PCI or coronary artery bypass grafting performed in cases of MI, unstable angina, stable angina, or silent ischemia. Malignant ventricular arrhythmia included ventricular fibrillation, ventricular tachycardia, and cardiac arrest. Heart failure was defined according to the European Society of Cardiology heart failure guidelines. LV dysfunction was defined as LVEF <50%. Lack of LVEF recovery was defined as failure to achieve LVEF ≥ 50% within 3 months after MI in patients with baseline LV dysfunction. In patients with multiple events, only the first occurrence after baseline was analyzed.

Data normality was evaluated using the Shapiro-Wilk test and auantile-quantile (QQ) plot visualization. Descriptive analyses included mean±standard deviation for normally distributed variables, median [interquartile range] for non-normally distributed variables, and frequency percentages for categorical variables. Comparative analyses between groups were conducted using the Student t test for normally distributed variables and the Mann-Whitney U test for nonnormally distributed variables, with categorical comparisons performed via the chi-square test or Fisher exact test. The cumulative incidence of outcomes was estimated using Kaplan-Meier analysis, with group differences assessed via the log-rank test. A Cox proportional regression model was performed to determine whether CMD was an independent predictor of the events, stratified by MetS categories. The relationships between CMD and LV dysfunction or lack of LVEF recovery were explored by multivariate logistic regression analyses. A 2-tailed P-value of <.05 was considered statistically significant. Statistical analysis was performed with in R 4.1.2 and SPSS 26.0.

This prospective study included a total of 497 STEMI patients (mean age 57.1 years, 85% male) who underwent both angio-IMR and baseline CMR analysis (figure 1). Among them, 316 (63.8%) patients had MetS, while 181 patients did not. The baseline demographic and clinical characteristics according to MetS and CMD category are presented in table 1. Door-to-wire time was longer in the CMD group than in the non-CMD group, for both patients with and without MetS. In patients with MetS, the CMD group had longer total ischemic time, higher high-density lipoprotein cholesterol levels, and a higher prevalence of current smoking and single-vessel disease. Among patients without MetS, the CMD group had higher systolic blood pressure and fasting blood glucose (table 1).

Figure 1.

Study flow diagram. CMD, coronary microvascular dysfunction; CMR, cardiac magnetic resonance; IMR, index of microcirculatory resistance; MetS, metabolic syndrome; STEMI, ST-elevation myocardial infarction.

(0.26MB).

The CMR indexes according to MetS and CMD category are presented in table S1. Compared with the non-CMD group, the CMD group exhibited lower LVEF, larger area at risk, larger infarct size, and a higher prevalence and greater extent of microvascular obstruction in patients both with and without MetS. Among patients with MetS, the CMD group had a higher prevalence and larger extent of intramyocardial hemorrhage at baseline. Among patients without MetS, the CMD group had larger LV end systolic volume index and LV mass index at baseline.

Over a follow-up time of 2.7±0.9 years (median 2.8 years), a total of 89 patients (17.9%) had MACCE. Patients with MetS had a higher rate of MACCE compared with patients without MetS (20.9% vs 12.7%; P=.014). The Kaplan-Meier curve for the primary endpoint is shown in figure 2.

Figure 2.

Kaplan-Meier survival curves of MACCE in STEMI patients stratified by MetS categories. MACCE, major adverse cardiovascular and cerebrovascular event; MetS, metabolic syndrome; STEMI, ST-elevation myocardial infarction.

(0.16MB).

The cumulative incidence of MACCE was higher in MetS patients with CMD than those without CMD (30.3% vs 18.4%; P=.034). In contrast, the cumulative incidence of MACCE was not significantly different between the CMD and non-CMD groups in non-MetS patients (12.6% vs 13.0%; P=.937) (table 2). The analysis of Kaplan-Meier curves also indicated a notably higher risk of MACCE in MetS patients with CMD (log-rank P=.023) (figure 3). Conversely, no significant difference was found between the CMD and non-CMD groups in non-MetS patients (log-rank P=.970) (figure 3).

Figure 3.

Kaplan-Meier survival curves of MACCE by CMD in STEMI patients (A) with MetS; and (B) without MetS according to angio-IMR. CMD, coronary microvascular dysfunction; MACCE, major adverse cardiovascular and cerebrovascular event; MetS, metabolic syndrome; STEMI, ST-segment elevation myocardial infarction.

(0.22MB).

Multivariable cox regression analyses also showed that CMD (angio-IMR> 40 U) was associated with a greater risk of MACCE in patients with MetS in all adjusted models, but not in patients without MetS, with the highest hazard ratio observed in model 3 (HR, 2.143; 95%CI, 1.164-3.944; P=.014) (table 3). This was true when angio-IMR was included as a continuous variable. Angio-IMR demonstrated a robust positive correlation with clinical risk in MetS patients, particularly in model 3 (HR, 1.022 per unit increase; 95%CI, 1.010-1.033; P <.001) (table 3). In contrast, no significant associations were observed in the non-MetS group for either angio-IMR as a continuous variable (all models P ≥ .346) or CMD (angio-IMR> 40U) (all models P ≥ .383), despite adjustments for covariates including age, sex, smoking status, prior stroke, ischemic time, and multivessel disease (table 3). The proportional hazards assumption was verified using Schoenfeld residuals, with no violations observed in either subgroup (figure S1 and figure S2). Additional sensitivity analyses further demonstrated that the prognostic association of CMD in patients with MetS remained significant when malignant ventricular arrhythmia or both malignant ventricular arrhythmia and stroke were excluded from the composite outcome, whereas no significant associations were observed in non-MetS patients (table S2).

Among patients with MetS, the CMD group had more severe LV enlargement (LV end systolic volume index, P=.032) and impairment in LV systolic dysfunction (LVEF, P=.031) at follow-up (table S3). Among patients without MetS, there was no difference in CMR indexes at follow-up between the CMD and non-CMD groups. The prevalence of LV dysfunction at baseline (56.1% vs 41.6%; P=.036) was significantly higher in MetS patients with CMD than in those with non-CMD (figure 4). At follow-up, MetS patients with CMD had a higher incidence of LV dysfunction (47.1% vs 23.6%; P=.015) and a lower incidence of absolute LEVF recovery (20.6% vs 30.6%; P=.029) than those without CMD (figure 4). In the multivariable regression analyses, CMD was also significantly associated with LV dysfunction (OR, 3.909; 95%CI, 1.330-11.489; P=.013) and lack of LVEF recovery (OR, 3.367; 95%CI, 1.099-10.318, P=.034) at follow-up, independently of baseline LVEF (table 4 and table 5).

Figure 4.

Rates of LV dysfunction and LVEF recovery by MetS and CMD. A: incidence of LV dysfunction at baseline in the 4 groups. B: incidence of LV dysfunction at the 3-month follow-up in the 4 groups. C: distribution of patients with normal baseline LVEF, LVEF recovery, and lack of LVEF recovery among the 4 groups. CMD, coronary microvascular dysfunction; LV, left ventricular; LVEF, left ventricular ejection fraction; MetS, metabolic syndrome.

(0.16MB).

Subgroup analyses based on individual metabolic risk factors revealed associations between CMD and clinical outcomes (figure 5). In patients with dysglycemia, CMD was associated with a significantly elevated risk of MACCE (HR, 2.150; 95%CI, 1.204-3.839; P=.010), whereas no such association was observed in normoglycemic individuals (HR, 1.000; 95%CI, 0.472-2.121; P=1.000). Similarly, among those with obesity, CMD conferred a marked increase in risk (HR, 2.662; 95%CI, 1.160-6.110; P=.021), while no significant effect was observed in normal BMI subgroups (HR, 1.195; 95%CI, 0.687-2.077; P=.529). In contrast, hypertriglyceridemia, hypertension, and low high-density lipoprotein cholesterol subgroups showed no statistically significant associations between CMD and outcomes across both strata (all P ≥ .098).

Figure 5.

Forest plot for major adverse cardiovascular and cerebrovascular events in CMD vs non-CMD stratified by different metabolic risk factors categories. 95%CI, 95% confidence interval; CMD, coronary microvascular dysfunction; HDL, high-density lipoproteins; HR, hazard ratio.

(0.26MB).

Our study has several limitations. First, CMR dropout remains an inherent limitation shared by other CMR studies in patients with STEMI. Second, other anthropometric indexes of obesity were not available, such as waist and hip circumference, which should be used as a supplement to BMI in future research. Third, as an observational study with a limited sample size and low event rate, causal inferences cannot be made, and some baseline differences could not be fully adjusted for, potentially affecting the generalizability of our findings. In particular, the low number of events in the non-MetS group may have reduced statistical power and contributed to the lack of observed association. Fourth, angio-IMR was measured at a single time point post-pPCI. The absence of serial CMD assessments during follow-up restricted our ability to analyze dynamic changes in microvascular function and their relationship with long-term outcomes.