Prognostic value of adenosine stress echocardiography in chronic coronary syndromes with preserved left ventricular ejection fraction

Abstract

Background

See More

The prognostic value of functional echocardiographic parameters for risk stratification in chronic coronary syndrome (CCS) remains incompletely understood. This study aimed to integrate resting and stress echocardiographic parameters to identify sensitive non-invasive predictors of major adverse cardiovascular events (MACEs).

Methods

A total of 754 CCS patients with a resting left ventricular ejection fraction (LVEF) ≥ 50% undergoing adenosine stress echocardiography were prospectively enrolled. Parameters including myocardial perfusion, coronary flow velocity reserve (CFVR), and myocardial work were assessed. Resting and stress values were compared within groups, while dynamic changes were analyzed between CFVR-normal and impaired subgroups. Cox regression was used to identify independent predictors.

Results

The incidence of MACEs was significantly higher in patients with impaired CFVR compared to those with normal CFVR (71.4% vs. 6.2%, P < 0.0001). After stress, the impaired CFVR group exhibited myocardial perfusion defects, mechanical dyssynchrony, and reduced myocardial work efficiency, in contrast to the normal CFVR group. Patients with impaired CFVR combined with regional wall motion and perfusion abnormalities had the worst prognosis. Multivariate Cox model integrating CFVR and stress-derived dynamic parameters demonstrated superior predictive performance for MACEs, significantly exceeding that of the stress-substitution and base models (C-index: 0.867 vs. 0.841 vs. 0.709).

Conclusion

In CCS patients with preserved LVEF, reduced CFVR indicates early myocardial dysfunction and predicts the worst prognosis when combined with wall motion and perfusion abnormalities. An integrated functional model combining CFVR and dynamic stress parameters enhances risk stratification for MACEs and supports individualized therapy.

Graphical Abstract

Highlights

1. In chronic coronary syndrome patients with preserved LvEF, reduced CFVR indicates early functional impairment.

2. Triple abnormality (impaired CFVR + regional wall motion abnormality + myocardial perfusion abnormality) predicts the highest MACEs risk.

3. Functional model integrating CFVR with stress-induced dynamic parameters enhances MACEs risk prediction.

Background

Most patients with chronic coronary syndrome (CCS) have a preserved left ventricular ejection fraction (LVEF). In the early stages, these patients often show normal resting electrocardiograms (ECG) and no regional wall motion abnormalities (RWMA) on echocardiography. Approximately 40% of patients with chest pain undergoing coronary angiography have no obstructive coronary lesions, likely due to the limited sensitivity of invasive angiography for diffuse atherosclerosis and coronary microvascular dysfunction [1]. The “pseudonormalization” seen on resting ECG and echocardiography frequently fails to detect underlying myocardial ischemia or microcirculatory dysfunction, thus underestimating actual cardiac function and anatomical lesions in CCS patients.

Stress echocardiography (SE) is a commonly used functional imaging tool for detecting inducible myocardial ischemia and stress-induced left ventricular dysfunction in patients with CCS [2, 3]. Its accuracy has been confirmed by multiple large studies and meta-analyses [4, 5]. However, traditional SE methods relying solely on RWMA have significant limitations [6]. They cannot reliably detect microvascular dysfunction or vasospastic angina, potentially leading to inaccurate prognostic stratification, particularly in CCS patients with preserved LVEF.

Adenosine SE induces coronary vasodilation via A2A receptor activation on vascular smooth muscle cells, enhancing microcirculatory flow and enabling evaluation of coronary vasodilatory capacity [6]. In cases with epicardial stenosis, adenosine induces the coronary steal phenomenon, allowing identification of ischemic regions through RWMA and reduced coronary flow velocity reserve (CFVR). Additionally, adenosine SE combined with 2D speckle-tracking imaging can quantify left ventricular myocardial strain and myocardial work parameters, providing more sensitive detection of early left ventricular dysfunction compared to traditional LVEF [7, 8]. However, current literature lacks systematic integration of diverse functional parameters from adenosine SE for predicting major adverse cardiovascular events (MACEs), and the incremental prognostic value of dynamic parameters remains unquantified.

This study employed a prospective cohort design to systematically evaluate echocardiographic parameters, including myocardial perfusion, coronary flow velocity (CFV), myocardial work, in CCS patients during rest and peak stress induced by adenosine SE. The objective was to establish a noninvasive risk stratification model based on the dynamic changes to stress.

Methods

Study population

This study consecutively enrolled CCS patients undergoing adenosine SE at our hospital from 2021 to 2023. Inclusion criteria were: (1) meeting current CCS diagnostic guidelines [9]; (2) resting LVEF ≥ 50%. The screening and enrollment process are detailed in Supplementary Figure S1. Baseline data, including demographics, cardiovascular risk factors, and medication use, were collected from hospital electronic records or standardized questionnaires. The study adhered to the Declaration of Helsinki and was approved by the Ethics Committee of Yan’an Hospital Affiliated to Kunming Medical University (approval No: 2023-179-01). All participants provided written informed consent.

Baseline transthoracic echocardiography

A GE Vivid E95 color Doppler ultrasound system (GE Vingmed Ultrasound, Horten, Norway) with an M5S phased-array probe (frequency 1.5–4.6 MHz) was used. Examinations were performed strictly following the American Society of Echocardiography guidelines [10]. Spectral and tissue Doppler imaging measured the peak velocity of tricuspid regurgitation (TRv), peak early diastolic blood flow velocities at the mitral valve orifice (E), and early diastolic motion velocities (e’) at the mitral annulus (interventricular septum and lateral wall). The average E/e’ ratio was calculated accordingly. Dynamic apical sectional images (apical four-, two-, and three-chamber views) covering at least three cardiac cycles were stored for offline analysis.

Adenosine SE protocol

A continuous intravenous infusion of adenosine (140 µg/kg/min, Penglai Nuokang Pharmaceutical, Shandong, China) was administered for 6 min. The ultrasound contrast agent SonoVue (Bracco Suisse SA, Geneva, Switzerland) was used to enhance imaging. SonoVue was prepared by dissolving 59 mg powder in 5.0 ml saline and shaking until fully dispersed. For each time, 1.5 ml of the solution was withdrawn and diluted with 5.0 ml saline.

The myocardial contrast echocardiography (MCE) protocol involved injecting the SonoVue solution into the cephalic vein at a uniform rate of 1 ml/10 s. After adequate myocardial filling of the left ventricle, a high-energy pulse was triggered to destroy microbubbles in the myocardium. The system automatically switched to low-energy (mechanical index = 0.1) real-time contrast imaging mode. Cine images of apical four-, two-, and three-chamber views, each containing 10 cardiac cycles post-pulse trigger, were acquired and stored for offline analysis.

Two-dimensional and MCE images at rest and during adenosine peak stress (5–6 min infusion) were collected. Criteria for terminating the adenosine SE test included severe angina with significant ST-segment elevation, marked blood pressure elevation (systolic > 180 mmHg or diastolic > 100 mmHg), severe arrhythmia, or reaching the study endpoint. During this study, 10 patients experienced premature atrial contractions, 3 patients experienced premature ventricular contractions, and 2 patients experienced first-degree atrioventricular block; all resolved after terminating the SE.

Coronary flow velocity reserve

CFV was measured at rest and 3 min into adenosine infusion using coronary Doppler mode. In a modified apical three-chamber view, pulsed-wave Doppler recorded diastolic peak flow velocity from the mid-to-distal left anterior descending artery. The color Doppler scale was set to 25 cm/s, and the probe position remained unchanged throughout imaging. CFVR was calculated using the formula: CFVR = CFV_Peak/CFV_Rest. An abnormal CFVR was defined as a value < 2.5¹¹.

Image and data analysis

Left ventricular volumes and LVEF were automatically calculated using speckle-tracking software by summing the end-systolic and end-diastolic stacks from apical four- and two-chamber views, with values expressed as the mean of the two views [12]. Stress-induced myocardial ischemia was defined by RWMA involving ≥ 2 contiguous segments with segmental scoring deterioration of at least one grade. Wall motion score index (WMSI) was evaluated using a 17-segment left ventricular model at both rest and peak stress. Segmental motion was graded on a 4-point scale: 1 (normal), 2 (hypokinesis), 3 (akinesis), and 4 (dyskinesis). Myocardial perfusion was assessed using real-time contrast-enhanced echocardiography following a high mechanical index “flash” impulse. Perfusion recovery was scored as follows: 1 = normal (replenishment within 4 s), 2 = delayed or reduced (replenishment between 4 and 10 s), and 3 = absent (no replenishment within 10 s) [13]. A perfusion score of 2 or 3 was defined as myocardial perfusion abnormality (MPA). The myocardial perfusion score index (MPSI) was calculated by dividing the sum of segmental scores by the number of segments.

Left ventricular function was assessed using automated function imaging. Apical four-chamber, two-chamber, and three-chamber views were acquired to automatically detect and track the endocardial border, allowing for the calculation of global longitudinal strain (GLS). To facilitate clinical interpretation, GLS values were reported as absolute values. Myocardial work (MW) parameters were derived using noninvasively measured brachial artery blood pressure and the pressure–strain loop methodology. These included global work index (GWI), global constructive work (GCW), global wasted work (GWW), and global work efficiency (GWE). Left ventricular contractile reserve (LVCR) was quantified by the stress-to-rest force ratio, calculated as systolic blood pressure divided by end-systolic volume. Dynamic parameters (Δ) were derived by subtracting resting from stress values.

Follow-up and major adverse cardiac events

Telephone and electronic medical record follow-up data were collected. The median follow-up was 571 days (range: 18–783 days). The primary endpoint was MACEs, defined as recurrent myocardial infarction, rehospitalization for heart failure or angina, new-onset ischemic chest pain with myocardial injury, malignant arrhythmias, cardiac arrest, cardiogenic shock, and all-cause death.

Statistical analysis

Statistical analyses were conducted using SPSS version 26.0 (IBM corporation, Armonk, NY, USA) and R version 3.5.2 (R Foundation for Statistical Computing, Vienna, Austria). Categorical variables are presented as counts (%), and continuous variables as mean ± standard deviation (if normally distributed) or median (interquartile range, IQR) (if not). Normally distributed variables were compared using independent-samples t-tests or paired t-tests, as appropriate. Non-normally distributed variables were compared using Mann-Whitney U tests. Categorical variables were analyzed by chi-square or McNemar’s tests, as applicable.

Univariate and multivariate Cox regression analyses assessed associations of demographic characteristics, clinical risk factors, and echocardiographic parameters (resting, stress-induced, and dynamic changes) with MACEs, presenting hazard ratios (HRs) and 95% confidence intervals (CIs). RWMA and MPA were analyzed as binary variables, while other parameters were continuous. Multivariate models were built as follows: Model 1 (base model) included demographics, clinical risk factors, and resting echocardiographic parameters; Model 2 (stress-substitution model) included the same demographic and clinical factors but replaced resting parameters with peak-stress parameters; Model 3 (functional-integration model) further incorporated dynamic parameters from resting and stress conditions. Variables were selected based on univariate significance (P < 0.05), with multicollinearity assessed via correlation coefficients (< 0.7) and variance inflation factors (VIF < 5) (Supplementary data online, Table S1). Model discrimination was assessed using log-likelihood values and the concordance index (C-index). Model discrimination was quantified by Harrell’s concordance index (C-index), and pairwise model comparisons used log-likelihood and DeLong’s test for C-index differences.

Kaplan-Meier analysis was performed to plot MACE-free survival curves, with differences assessed using log-rank tests. Predictive performance of echocardiographic parameters in the optimal model for MACEs was evaluated by ROC curve analysis, calculating area under the curve (AUC) with 95% CI, and determining optimal cutoffs using the Youden index. Pairwise comparisons were performed with a corrected alpha level to minimize the risk of Type I error. All tests were two-sided, with P < 0.05 indicating statistical significance.

Results

Clinical characteristics of the study population

Of the 754 CCS patients, 62% were male, with a median age of 59.5 (53–67) years. Among them, 444 (59%) had hypertension, 110 (15%) had hyperlipidemia, and 129 (17%) had diabetes. Additionally, 47 patients (6%) had a history of non-coronary heart diseases, including small atrial septal defect/patent foramen ovale, mild-to-moderate valvular regurgitation, and paroxysmal supraventricular tachycardia. Baseline characteristics of the study population are presented in Table 1.

Table 1 Clinical characteristics of study population (n = 754)

Echocardiographic characteristics at rest and peak stress

Patients were divided into a normal CFVR group (CFVR ≥ 2.5, n = 390) and an impaired CFVR group (CFVR < 2.5, n = 364). Both groups showed increased heart rate and decreased blood pressure after stress. Although LVEF, GLS, and E/e’ values increased from rest to peak stress in both groups, significant intergroup differences were observed in wall motion and myocardial perfusion. The impaired CFVR group demonstrated higher rates of RWMA and MPA, with increased WMSI and MPSI scores, whereas the normal CFVR group reductions in these parameters (all P < 0.01). For myocardial work, the normal CFVR group exhibited increased GWI and GWE and decreased GWW post-stress, while the impaired CFVR group showed reductions in GWI, GCW, and GWE (Table 2). Figure 1A and B demonstrate characteristic functional differences between cases with normal and impaired CFVR by comparing CFV, GLS, MW indices, and myocardial perfusion at rest and during peak stress.

Table 2 Echocardiographic characteristics comparison at rest and peak stress

Fig. 1 

A Representative case with normal CFVR: CFVR > 2.5 with normal GLS, and myocardial work, and myocardial perfusion. Upper panels: Resting images; Lower panels: Peak stress images. B Representative case with impaired CFVR: CFVR < 2.5, normal GLS, with reduced GWI, GCW and GWE, and increased GWW after stress. Myocardial perfusion abnormality observed (arrow). Coronary angiography revealed 90% stenosis of the right coronary artery. Upper panels: Resting images; Lower panels: Peak stress images

Comparison of echocardiographic parameters between normal vs. Impaired CFVR groups

Compared with the normal CFVR group (median CFVR: 2.86), the impaired CFVR group (median CFVR: 2.00) demonstrated no significant difference in LVCR (P = 0.194), but had smaller increases in ΔLVEF and ΔGLS, along with significantly greater increases in ΔE/e’ and ΔTRv (all P < 0.05). In myocardial work parameters, the impaired CFVR group showed no significant change in ΔGWE and a blunted reduction in ΔGWW, while both ΔGWI and ΔGCW decreased significantly. In contrast, the normal CFVR group achieved significant increases in ΔGWI and ΔGWE (both P < 0.01). Additionally, ΔWMSI and ΔMPSI were markedly elevated in the impaired CFVR group (both P < 0.01) (Table 3).

Table 3 Comparison of echocardiographic parameters: normal vs. Impaired CFVR

Outcomes

The median follow-up duration was 571 days (IQR: 18–783 days). Endpoint events occurred in 284 patients (38%). In the impaired CFVR group, 260 patients (71.4%) experienced MACEs, whereas in the normal CFVR group, 366 patients (93.8%) remained event-free. The difference in the incidence of MACEs between the two groups was significant (P < 0.0001) (Supplementary data online, Table S2).

Kaplan-Meier event-free survival curves

Based on CFVR, regional wall motion, and myocardial perfusion status, patients were stratified into four groups:

• Group A: CFVR ≥ 2.5 with normal wall motion and myocardial perfusion;

• Group B: CFVR < 2.5 with normal wall motion and myocardial perfusion;

• Group C: CFVR < 2.5 with either RWMA or MPA;

• Group D: CFVR < 2.5 with both RWMA and MPA.

Kaplan–Meier survival analysis revealed a stepwise decline in MACEs-free survival across the four groups (log-rank P < 0.001). with the poorest prognosis observed in patients with reduced CFVR accompanied by both RWMA and MPA. (Fig. 2)

Fig. 2

Kaplan–Meier curves for MACEs by stress echocardiography status. Group A: CFVR ≥ 2.5 with normal wall motion and myocardial perfusion; Group B: CFVR < 2.5 with normal wall motion and myocardial perfusion; Group C: CFVR < 2.5 with either RWMA or MPA; Group D: CFVR < 2.5 with both RWMA and MPA. MACEs, major adverse cardiac events. RWMA, regional wall motion abnormality; MPA, myocardial perfusion abnormality

Association between echocardiographic parameters and maces

Univariate Cox analysis indicated that age, male sex, hypertension, diabetes, and several echocardiographic parameters at rest and under stress were significantly associated with MACEs (Supplementary data online, Table S3). Three hierarchical multivariate Cox regression models were constructed (Table 4): Model 1 (base model) included demographic characteristics, clinical risk factors, and resting echocardiographic parameters; Model 2 (stress-substitution model) replaced resting parameters with peak-stress parameters while maintaining the same demographic and clinical factors; Model 3 (functional integration model) further integrated dynamic parameters derived from stress and resting states. In Model 1, age, male sex, CFV, and MPA were independent risk factors, while LvEF had a protective effect. In Model 2, hypertension, diabetes, and MPA were significantly associated with MACEs, whereas CFV and GLS became a protective factor. Model 3 showed hypertension, ΔTRv,ΔGWW, and LVCR as risk factors and CFVR and ΔGLS as a protective factor. Model discrimination, quantified by the C-index, showed progressively improvement across models: 0.709 for Model 1, 0.841 for Model 2, and 0.867 for Model 3. Pairwise comparisons using log-likelihood ratio tests and DeLong’s test confirmed the significantly superior predictive performance of Model 3 over both Model 2 and Model 1 (all P < 0.001) (Fig. 3).

Table 4 Multivariable Cox regression models for major adverse cardiac events

Fig. 3

Model comparison based on log-likelihood ratio tests (left) and C-index values (right). Model 3 demonstrated significantly improved predictive performance compared to Models 2 and 1

ROC curves of echocardiographic parameters for maces

ROC analysis of the optimal model (functional integration model) with Bonferroni-adjusted α for multiple comparisons revealed strong predictive value of CFVR (AUC = 0.883) and ΔGLS (AUC = 0.720) for MACEs. ΔTRv (AUC = 0.695) and ΔGWW (AUC = 0.634) were also independently associated with MACE risk (all P < 0.0001), while LVCR (AUC = 0.589) demonstrated relatively limited predictive ability (Fig. 4). The AUC, 95% CI, optimum cutoff points, sensitivity, and specificity are detailed in Supplementary Table S4.

Fig. 4

ROC curves of echocardiographic parameters in Model 3 for predicting MACEs. CFVR, coronary flow velocity reserve; GLS, global longitudinal strain; TRv, tricuspid regurgitation velocity; GWW, global wasted work; LVCR, left ventricular contractile reserve; Δ, delta peak-rest; MACEs, major adverse cardiac events

Discussion

The main findings of this study were as follows: (1) Despite preserved global contractile reserve, patients with impaired CFVR demonstrated early functional deterioration, including diastolic dysfunction, mechanical dyssynchrony, and reduced myocardial energy efficiency. (2) The combination of reduced CFVR, regional wall motion abnormalities, and impaired myocardial perfusion identified patients with the poorest survival and significantly elevated MACEs risk. (3) The functional integration model demonstrated the highest predictive performance, identifying CFVR, ΔTRv, ΔGWW, LVCR, and ΔGLS as independent prognostic factors, highlighting the clinical value of comprehensive stress-derived parameters in risk stratification.

Stress imaging induces myocardial oxygen supply-demand imbalance to trigger the ischemic cascade [14]enabling detection of early left ventricular functional abnormalities, such as regional wall motion or impaired myocardial perfusion, prior to global ventricular dysfunction, ECG abnormalities, or angina. Although exercise SE is widely used as a first-line screening method, this study employed a high-dose adenosine stress protocol, which offers advantages such as more stable CFVR measurements, independence from achieving peak heart rate, and shorter examination duration.

RWMA are clinically validated markers for detecting myocardial ischemia during SE [15]demonstrating high sensitivity and specificity for obstructive coronary artery disease. This diagnostic capability arises from transmural progression of ischemia—from subendocardial to subepicardial layers. Transmural ischemia characteristically manifests as RWMA [16]. However, RWMA has limited value in identifying ischemia caused by non-obstructive coronary pathologies such as microvascular dysfunction, where circumferential or diffuse myocardial involvement typically fails to produce detectable RWMA [15].

Previous studies established resting GLS as an independent predictor of MACEs [17, 18]with absolute values < 15% significantly associated with increased all-cause mortality and cardiac endpoints [17, 19]. LVCR also demonstrates risk-stratification value [20]. In this cohort of CCS patients with preserved LVEF, stress-induced increases in LVEF, improved GLS (> 15%), and normal LVCR (< 1.1) collectively indicated preserved global contractile reserve. Critically, multivariable Cox regression identified dynamic functional indices as superior prognostic markers: ΔGLS independently protected against MACEs, whereas LVCR constituted an independent risk factor. These results establish that dynamic parameters provide incremental prognostic value beyond traditional resting or single-point stress measures, enhancing risk stratification in CCS.

Myocardial perfusion abnormality

The prognostic significance of myocardial perfusion is well-established through multiple large-scale investigations [6, 21]. A meta-analysis of 5953 subjects demonstrated that stress-induced MPA significantly outperformed RWMA and LVEF in predicting MACEs (HR 4.75 vs. 2.39 vs. 1.92), showing independent associations with mortality, non-fatal myocardial infarction, and revascularization requirements [22].

In our cohort of patients with CCS, both resting and stress-induced MPA independently predicted MACEs, with significantly higher risk during stress (HR 5.146 vs. 2.535). Crucially, ΔMPSI increased in patients with impaired CFVR but decreased in those with normal CFVR, indicating stress-induced compromise of coronary perfusion reserve. This abnormal ΔMPSI elevation, particularly pronounced in patients with limited microcirculatory compensatory capacity, likely reflects endothelial-dependent vasodilator dysfunction or coronary steal phenomenon, constituting a biomarker of elevated cardiovascular risk.

Coronary flow velocity reserve

CFVR is established as a robust prognostic marker in patients with CCS [23]. With a measurement success rate of up to 95% under standardized vasodilator stress protocols, its clinical applicability is well supported [6]. Evidence from multicenter studies has confirmed its independent association with all-cause mortality across a range of cardiac conditions [24]regardless of the severity of epicardial coronary stenosis [25]. Crucially, CFVR demonstrated complementary risk stratification to regional wall motion assessment: reduced CFVR identified high-risk subgroups even among patients without RWMA [26]. A multicenter study of 7,576 CCS patients with LVEF ≥ 50% showed that rest CFV ≥ 32 cm/s combined with CFVR ≤ 2.0 identifies patients at high risk of MACEs [27].

The present study reinforces these findings by demonstrating the superior predictive performance of CFVR for MACEs (AUC = 0.88, 95%CI: 0.86–0.91). The functional integration model incorporating CFVR achieved optimal diagnostic accuracy (C-index 0.867), outperforming the other models. These results highlight CFVR’s central role in clinical risk stratification, particularly in complex cases dominated by microvascular disease or where anti-anginal medications confound RWMA interpretation.

Myocardial work

MW analysis integrates dynamic left ventricular pressure-strain interactions, mitigating the load-dependency limitations of GLS to provide comprehensive functional assessment in CCS [28]. Current evidence indicates MW parameters outperform GLS in detecting significant coronary artery disease (CAD) [29]with segmental MW demonstrating vessel-specific responses to territorial stenoses [30]. The clinical utility of MW is further enhanced under stress conditions, as evidenced by an exercise stress echocardiography study of 85 angina patients where combined peak GWE (AUC = 0.836) and recovery-phase GWW (AUC = 0.856) surpassed individual parameters in identifying significant CAD [31].

This study further extends MW applications to vasodilator adenosine SE. Patients with normal CFVR exhibited improved myocardial work efficiency, whereas those with impaired CFVR showed reductions in GWI, GCW, and GWE. This pattern indicates compromised myocardial energy utilization under diminished coronary flow reserve. Critically, ΔGWW emerged as an independent prognostic factor in multivariable modeling, indicating that accumulated wasted work reflects progressive microvascular ischemia. These findings establishes ΔGWW as a novel biomarker for refined risk stratification in CCS.

Incremental prognostic value of multiparametric integration

While conventional single functional parameters have established roles in cardiovascular risk stratification [17]their predictive capacity remains limited in CCS patients with preserved LVEF [32,33,34]. Previous studies indicate significant prognostic improvement when SE evolves from isolated wall motion assessment to integrated hemodynamic and metabolic profiling [35]. The combined CFVR and LvCR abnormalities increases the 3-year MACEs risk to 63%, substantially higher than with single-parameter abnormalities (42% for CFVR abnormality, 19% for LVCR abnormality) or normal results (10%)³⁶.

In this study, a stepwise modeling strategy validated the incremental prognostic value of peak-stress and dynamic functional parameter integration. The stress-substitution model demonstrated enhanced predictive performance versus the base model (C-index 0.841 vs. 0.709), confirming the independent significance of stress-induced myocardial strain and perfusion abnormalities. Subsequent integration of ΔGLS, ΔGWW, CFVR, and LVCR achieved optimal discrimination (C-index 0.867). Critically, the independent prognostic contributions of ΔGWW and CFVR underscore microcirculatory dysfunction-induced impairment in myocardial energy utilization efficiency while revealing inherent limitations of conventional wall motion analysis in detecting high-risk subgroups. This functional integration model demonstrates substantial clinical utility for CCS risk stratification, providing a robust foundation for personalized diagnostic and therapeutic strategies.

Limitations

This study has several limitations. Firstly, as a single-center study, there may be selection bias. Clinical decisions regarding invasive procedures for inpatients might also have influenced outcomes, thus limiting the external validity of the findings. Secondly, the follow-up period was relatively short (median 571 days). Follow-up data relied mainly on telephone interviews and electronic medical records, posing a risk of information omission or recall bias. Additionally, the study population had a relatively high proportion of comorbidities, and many patients were taking medications. These medications could indirectly influence functional parameters by improving myocardial oxygen supply-demand balance. Due to ethical constraints, patients could not discontinue these medications, making it difficult to fully exclude drug-related confounding effects [37].

Echocardiographic parameters carry inherent technical considerations. RWMA and MPA require qualitative interpretation dependent on operator expertise, introducing subjective bias risk. CFVR, calculated as stress-to-rest velocity ratio, lacks absolute flow volume and microcirculatory resistance data—limiting precision in microvascular dysfunction assessment. Parameters like CFVR, ΔGLS, ΔGWW, and LVCR exhibit measurement variability from image quality limitations, frame rate constraints, and operator dependency. Clinical interpretation thus demands rigorous evaluation of technical biases and confounders. Future multicenter prospective studies implementing AI-assisted quantification could enhance measurement reliability.

Conclusions

In CCS patients with preserved LVEF, impaired CFVR signifies early cardiac dysfunction despite maintained global contractile reserve. The most adverse prognosis manifests when reduced CFVR coexists with regional wall motion and myocardial perfusion abnormalities. The functional integration model incorporating CFVR and dynamic stress parameters demonstrates significantly enhanced risk stratification performance.