Coronary artery disease (CAD) is associated with high morbidity and mortality rates worldwide (1). Prognostic assessment is critical when deciding to treat patients with CAD and formulating prevention strategies. The main method used for the stratification of risk among patients with CAD is the evaluation of stress-induced myocardial ischemia, often by single-photon emission computed tomography (SPECT) myocardial perfusion imaging (MPI) (2). Essentially, gated MPI (G-MPI) enables the simultaneous assessment of the distribution of myocardial perfusion and cardiac function. Risk stratification can be enhanced by applying multiple parameters acquired by G-MPI, including myocardial perfusion data and functional information. A range of key factors, such as enlarged ventricular volume, the presence of transient ischemic dilatation (TID), and, in particular, reduced left ventricular ejection fraction (LVEF), have been identified as independent risk factors for adverse outcomes in patients with CAD (3).

LVEF is the preferred variable for evaluating LV systolic function (4). Furthermore, a reduction in LVEF reserve (ΔLVEF), defined as LVEF_Stress minus LVEF_Rest (5), has been associated with ischemic contractile dysfunction (6, 7). Previous studies utilizing ⁸²Rb positron emission tomography (PET) myocardial perfusion imaging demonstrated that ΔLVEF represented an independent predictor of major adverse cardiac events (MACEs) (5, 8). Nevertheless, the prognostic value of ΔLVEF, as measured by SPECT G-MPI (9–11) has yet to be fully evaluated. Besides, most studies did not specifically focus on patients with cardiac dysfunction (10–12). Furthermore, the incremental prognostic value of an abnormal ΔLVEF in patients with reduced LVEF_Stress has yet to be investigated. In addition, research has shown that the extent and severity of myocardial ischemia can both influence the prognosis and a large area of ischemia (>10%/LV) is considered to be a key indicator of revascularization for patients with CAD (13). There is a significant paucity of data relating to the prognostic value of ΔLVEF in patients with varying degrees of myocardial ischemia, especially in patients with no or mild myocardial ischemia.

Therefore, this study aimed to evaluate the prognostic value of ΔLVEF, as determined by SPECT G-MPI, in patients diagnosed with CAD and in a high-risk group of patients with left ventricular dysfunction (LVEF_Stress < 60%). In addition, we analysed the prognostic value of ΔLVEF in patients with different degrees of myocardial ischemia.

This study aimed to evaluate the prognostic value of ΔLVEF, as determined by SPECT G-MPI in patients with CAD, to predict MACEs. Our results indicated that in patients with a LVEF_Stress < 60%, an ΔLVEF ≤ 0% was identified as independent predictors of MACEs by multivariate Cox regression analysis. Furthermore, in patients with no or mild myocardial ischemia, the incidence of MACEs in the ΔLVEF ≤ 0% group was significantly higher than in the ΔLVEF > 0% group. Moreover, adding ΔLVEF to the traditional perfusion and functional variables of MPI significantly improved the discriminatory power to predict MACEs. Our results were generally consistent when left ventricular systolic dysfunction was defined as LVEF_Rest < 55%.

LVEF has been a key variable for the diagnosis and management of heart failure. In our study, we specifically focussed on LVEF_Stress < 60% because the latest recommendations by the British Society of Echocardiography (14) and the American Society of Nuclear Cardiology (24) state that the cut-off value for a “normal” LVEF is 55%. However, differences have been identified in terms of sex, age, and ethnicity. For example, there is a clear difference in LVEF between Europeans and Asians. The predicted values for Europeans are known to be significantly lower than those for East Asians. Specifically, for both sexes (at the age of 50 years), the lower reference value of LVEF for Europeans was 6% lower than that for East Asians. Furthermore, tenfold more Europeans than East Asians were found to have an LVEF < 50% (25). Unfortunately, while the LVEF criteria are applicable and appropriate for European populations, there is a significant scarcity of available data relating to LV function parameters acquired by gated MPI in the Chinese population. In addition, we cannot ignore the wide limits of agreement between echocardiography and SPECT G-MPI when determining LVEF (26). In routine clinical practice, we recommend monitoring borderline LVEF to avoid delay or missing high-risk patients. Our recent study (15) provided insights into the normal reference values of LVEF_Stress when measured by D-SPECT G-MPI in both women and men, which were 70 ± 8% and 68 ± 7%, respectively. Therefore, our centre gives considerable attention to patients with LVEF_Stress < 60%. Meanwhile, the results were generally consistent when left ventricular systolic dysfunction was defined as LVEF_Rest < 55%, but not when defined as LVEF_Stress < 55%. Indeed, in some centres, the stress-only strategy, or stress-first strategy, has been implemented to reduce costs and enhance the efficacy of testing (27, 28). Therefore, an appropriate expansion of the criteria for LVEF_Stress reduction aligns with clinical practice.

Previous evidence showed that a reduced ΔLVEF, as determined by 82Rb PET MPI, serves as a marker for ischemic contractile dysfunction (7) and is associated with an increased risk of cardiac events (5) and all-cause mortality (18). However, the existing literature describes inconsistent findings concerning the predictive significance of ΔLVEF when determined by SPECT MPI (9, 10, 12). In a previous study, Smith et al. (12) demonstrated that an abnormal LVEF reserve was not associated with an increased risk of the primary outcome. One possible explanation for this difference is that most patients in the study reported by Smith et al. (12) underwent a single-day protocol. In contrast, a two-day protocol was used in the present study as per our routine clinical practice. In the single-day protocol, the rest examination was performed approximately three hours after the stress examination, possibly leading to an underestimation of the alteration in LVEF, particularly in patients with severe ischaemia who may have experienced prolonged stunning. In addition, our landmark analysis revealed the effect of ΔLVEF on long-term prognosis. Specifically, in our cohort of patients with coronary stenosis and left ventricular systolic dysfunction, 23 MACEs were observed beyond 2 years, a large proportion (78.3%) of whom underwent incomplete revascularization (n = 2) and conservative strategies (n = 16). BARI-2D (29) found in high-risk patients, including those with reduced LVEF and extensive coronary disease, the five-year risk of death/MI/stroke were significantly lower among those undergoing revascularizations when compared with the group of medical therapy alone. In particular, the survival curve showed a significant increase in the difference in event rates after 2 years. Similarly, STICH trial (30) reported a significant benefit began to accrue after 2 years when comparing CABG and medical therapy in patients with heart failure. Our results strongly correlated with the above reports. We speculate that the absence of LVEF reserve may indicate a declining cardiac reserve, and that coronary artery stenosis and progressive myocardial ischemia may contribute to this poor prognosis in the later stages.

In a previous study, Gomez et al. (9) defined an abnormal ΔLVEF as a reduction of <5% in LVEF in post-stress images. This criterion was derived from a previous study (31) that proposed a 5% threshold for ΔLVEF when distinguishing between normal and abnormal responses. The study demonstrated that a ΔLVEF of 5% provided the highest diagnostic accuracy (sensitivity 52%, specificity 83%) for detecting multivessel CAD. Nevertheless, the most extensive cohort study to date (10), featuring 10,275 patients who underwent SPECT-MPI, revealed that an increase of 1% in LVEF reserve was significantly and independently associated with a lower incidence of MACEs, including cardiac death and myocardial infarction [HR: 0.98; 95% CI: 0.97, 0.99, P = 0.003]. Thus, additional clarification is needed to enable a more significant prognostic capability for patient outcomes. Within our present cohort, only 9.2% (n = 24) of patients exhibited an ΔLVEF of ≥5%, thus indicating that an ΔLVEF of ≤0%, rather than an ΔLVEF of ≤5%, represents a crucial and autonomous prognostic marker, thereby aligning with recent research (11, 18, 19). However, the prognostic value of ΔLVEF, as determined by SPECT MPI, has not been reported in a high-risk cohort with a reduced LVEF. This study is the first to report that an ΔLVEF ≤ 0% was an independent predictor of MACEs in patients with a LVEF_Stress < 60%. This finding provides a valuable point of reference for guiding future clinical practice.

Previous research established the importance of myocardial ischemia for determining therapeutic strategies. Patients with no to mild ischemia were categorised as low risk, for whom a conservative treatment approach was considered to be appropriate. In contrast, patients with moderate to severe ischemia were recommended for revascularization to improve their prognosis (32, 33). To the best of our knowledge, there is a significant scarcity of data relating to the prognosis of ΔLVEF in patients with varying degrees of myocardial ischemia. Smith et al. (12) previously performed subgroup analysis for patients with large areas of ischemia (≥10%/LV) and determined no significant difference in the incidence of primary outcomes compared to those with and without LVEF reserve. These findings are consistent with those arising from our present analysis. Unfortunately, the study lacked data on patients with no to mild ischemia. Our results suggest, for the first time, that the combination of ΔLVEF with the extent of myocardial ischemia could enhance risk stratification in patients with CAD. Notably, patients with no to mild myocardial ischemia were considered to have a favourable prognosis, whereas those with an ΔLVEF ≤ 0% exhibited a relatively high risk of MACEs.

A large area of myocardial ischemia has been confirmed to be associated with poor outcomes in CAD patients. Its prognostic effect is very strong and significant. In our cohort, a total of 44 MACEs were observed in patients with moderate to severe myocardial ischemia, with 75% (n = 33) occurring within 2 years. Our results reveal that the influence of the ischemia on prognosis was significantly greater than that of ΔLVEF in a short term. The influence of ΔLVEF on outcomes has gradually become more apparent with the progression of the disease. In fact, the mechanism of the prognostic significance of ΔLVEF is not very clear at present. We observed that the ΔLVEF ≤ 0% group exhibited a higher LVEF_Rest than the ΔLVEF > 0% group, whereas LVEF_Stress was higher in the ΔLVEF > 0% group. Furthermore, the 79.5% of impaired LVEF_Stress (≤50%) (23) was included in the ΔLVEF ≤ 0% group (vs. ΔLVEF > 0% group, P = 0.005). In contrast, a slightly higher proportion of individuals with supra-normal left ventricular ejection fraction (snLVEF) (LVEF_Rest ≥ 65%) were found in the ΔLVEF ≤ 0% group (n = 9, P = 0.050), compared to ΔLVEF > 0% group (n = 0). The snLVEF is considered to be associated with a poor prognosis (34), but the mechanism is unclear. We speculate that the combination of potential functional abnormalities in the resting state and impaired cardiac reserve, which presents a poor response to stress, may lead to a reduced ΔLVEF and posing a risk of long-term poor prognosis. Further research into this potential relationship is needed.

To our knowledge, only one previous study investigated the incremental value of ΔLVEF for predicting MACEs beyond the conventional variables of MPI. Otaki et al. (11) recruited 151 patients undergoing same-day rest/stress SPECT G-MPI. Early stress imaging was initiated 2 min after the injection of regadenoson, followed by late-stress acquisition. This study demonstrated that adding ΔLVEF during early stress enhanced the combined model of age, a prior history of PCI, and TPD (P < 0.001). The annualised MACEs rates during the late-stress period exhibited variances between patients with an ΔLVEF < 0% (6.7%) and an ΔLVEF ≥ 0% (4.9%), although these differences were not statistically significant. However, the sample size of this previous study was limited and focused explicitly on preserved LVEF_Stress, unlike our current study. Furthermore, Otaki et al. did not analyse the traditional parameters of MPI, except for TPD. It is widely acknowledged that larger perfusion defects, reduced ejection fraction, and larger ventricular volume predict adverse cardiac events (10, 35). Our current findings concur with these earlier findings. Based on our current findings, we emphasize that in patients with left ventricular dysfunction, both stress and resting MPI parameters, including TPD, ischemia, and LVEF, particularly ΔLVEF, may provide valuable assistance for the further risk stratification of patients with CAD.

The European Society of Cardiology guidelines (13) published recently for managing chronic coronary syndromes (CCS), guide clinicians in choosing imaging techniques (36). Both functional and anatomical aspects must be considered in patients with suspected CCS, and the importance of non-invasive imaging for selecting patients to be referred for invasive angiography has been emphasized. In particular, functional assessment may be crucial for identifying the mechanisms behind myocardial ischemia and, eventually, angina, thus guiding symptomatic treatment (37). Speckle tracking echocardiography (STE) is a reliable and widely used imaging technique of recognized clinical value in several settings. This method uses the motion of ultrasound backscatter speckles within echocardiographic images to derive myocardial velocities and deformation parameters (38). Notably, global longitudinal strain (GLS) is considered an earlier marker of myocardial damage and predicts mortality in patients with CCS independently of LVEF (39). The myocardial deformation imaging might reveal subtle abnormalities that can be attributed to clinically relevant ischemic or ischemic memory (40). This ischemic memory may be considered relevant to myocardial stunning and the reduction of post-stress LVEF (41). Integrating multiple imaging modalities and attempting to reveal the pathophysiological mechanisms is an important direction for future research.

Some studies support the notion that the presence of TID can specifically indicate extensive or severe coronary artery disease (42). However, this study found that TID was not an independent predictor for MACEs, thus aligning with a large cohort study previously conducted by Kattoor et al. (10) who found that the prognostic value of ΔLVEF was higher than that of TID. The pathophysiology of TID remains controversial (43, 44), although the predominant hypothesis is that TID originates from either diffuse subendocardial hypoperfusion leading to an apparent increase in LV endocardial cavity size and/or stress-induced LV dysfunction (3). Although investigating a specific group of patients with LV dysfunction may influence the prognostic value of TID, it is noteworthy that we identified clear differences between the ΔLVEF groups for ΔLVESV and TIDESV but not for ΔLVEDV or TIDEDV, thus indicating an association between a reduction in post-stress LVEF and left ventricular systolic dysfunction.

Our research is subject to several limitations that need to be considered. First, owing to its retrospective nature and the fact that this was a single-centre study with a relatively small sample size, there is potential for selection bias. To reduce the waiting list time for MPI, we did not perform rest studies in patients with normal stress-gated MPI in our laboratory. Therefore, even if more than 3,000 MPI studies were conducted per year, the number of populations in the current study was limited. Another limitation was the acquisition of gated MPI, which was performed after 60–90 min according to different stress or rest states. This implies that the acquisition of LVEF_Stress by SPECT was not derived during peak stress. However, we confirmed the predictive value of ΔLVEF in patients with a LVEF_Stress < 60%, particularly in those with no or mild myocardial ischemia. Our findings emphasise that combining perfusion and cardiac function parameters may enhance risk stratification.

In this pilot study, we found that when determined by SPECT G-MPI, ΔLVEF was independently associated with MACEs in CAD patients with LVEF_Stress < 60%, enhancing risk stratification for MACEs. Patients with no to mild myocardial ischemia were considered to have a favourable prognosis, whereas those with an ΔLVEF ≤ 0% exhibited a relatively high risk of MACEs. This is a pilot study with a small sample size, and further investigation and validation are needed.