Predictors of Long-Term Mortality in Patients with Stable Angina Pectoris and Coronary Slow Flow

Background and Objectives: Coronary slow flow (CSF) is an angiographic phenomenon characterized by the slow progression of an injected contrast agent during diagnostic coronary angiography in the absence of significant stenosis. Although CSF is a common angiographic finding, the long-term outcomes and mortality rates are still unknown. This study aimed to investigate the underlying causes of mortality over a 10-year period in patients diagnosed with stable angina pectoris (SAP) and CSF. Materials and Methods: This study included patients with SAP who underwent coronary angiography from 1 January 2012 to 31 December 2012. All patients displayed CSF despite having angiographically normal coronary arteries. Hypertension (HT), diabetes mellitus (DM), hyperlipidaemia, medication compliance, comorbidities, and laboratory data were recorded at the time of angiography. Thrombolysis in myocardial infarction (TIMI) frame count (TFC) was calculated for each patient. The cardiovascular (CV) and non-CV causes of long-term mortality were assessed. Results: A total of 137 patients with CSF (93 males; mean age: 52.2 ± 9.36 years) were included in this study. Twenty-one patients (15.3%) died within 10 years of follow-up. Nine (7.2%) and 12 (9.4%) patients died of non-CV and CV causes, respectively. Total mortality in patients with CSF was associated with age, HT, discontinuation of medications, and high-density lipoprotein cholesterol (HDL-C) levels. The mean TFC was associated with CV mortality. Conclusion: Patients with CSF exhibited a notable increase in cardiovascular-related and overall mortality rates after 10 years of follow-up. HT, discontinuation of medications, HDL-C levels, and mean TFC were associated with mortality in patients with CSF.


Introduction
Coronary slow flow (CSF) was first described by Tambe et al. [1]. CSF is characterized by the slow progression of an injected contrast agent during coronary angiography in the absence of significant stenosis. It is observed in 1-7% of coronary angiograms and is more common among young people and smokers [2,3]. Particularly, CSF is typically observed in patients who undergo angiography for acute coronary syndrome, usually unstable angina. Nearly two-thirds of patients with CSF presented with an acute coronary syndrome [2]; however, a small percentage (8%) of patients may present with acute myocardial infarction [4]. CSF increases the risk of recurrent chest pain, recurrent hospitalizations, recurrent cardiac catheterizations, life-threatening arrhythmias (e.g., Torsades de Pointes), and sudden cardiac death [5,6]. CSF can also cause myocardial ischemia and recurrent acute coronary syndrome (ACS) [2,4].
Despite being a well-known angiographic finding, the exact pathogenic mechanisms of CSF are not fully understood; however, endothelial dysfunction, microvascular abnormalities, and atherosclerosis have been suggested to play a role in the pathogenesis of CSF [7]. Long-term outcomes and mortality in patients with CSF remain to be elucidated.

Study Population
The study enrolled patients who had stable angina pectoris (SAP) and CSF. All coronary angiographies performed between 1 January 2012 and 31 December 2012 were reviewed for the presence of CSF. All patients who underwent coronary angiography for SAP at the Siyami Ersek Thoracic and Cardiovascular Surgery Centre from 1 January 2012 through 31 December 2012 who fulfilled the criteria for a diagnosis of CSF were included. CSF was diagnosed according to the criteria described below [3]. All patients had normal coronary arteries on angiography. Past medical histories, physical examination findings, medications, electrocardiograms (ECG), and echocardiograms were obtained. Patients with ACS were excluded.
A total of 137 patients (93 males) with angiographically normal coronary arteries and CSF were included in this study. Patients with visible plaques on angiography were excluded. Other exclusion criteria were a history of valvular heart disease, congenital heart disease, coronary and peripheral artery disease, atrial fibrillation, aortic aneurysm, revascularization, coronary artery ectasia, heart failure, ACS, pulmonary embolism, chronic obstructive pulmonary disease, acute or chronic infections, cancer, autoimmune or inflammatory diseases, thyroid disease, the administration of medication with anti-inflammatory properties, hepatic or renal dysfunction, left ventricular dysfunction, and left ventricular hypertrophy on echocardiography.
Electronic medical records were used to obtain data on symptoms, physical examination findings, laboratory test results, recurrent hospitalizations, and medications, as well as ECG and echocardiography reports. Ten-year follow-up status was determined by inviting patients to the hospital by phone. Follow-up medical histories were obtained, and physical examinations and ECGs were performed. Compliance with physician follow-up and medications was determined. Information regarding deceased patients was obtained from their relatives. Comorbidities and causes of death were extracted from the hospital's electronic medical records, the national health system database, and death certificates.
Demographic data such as age and sex, as well as medical history data pertaining to patients' histories of hypertension (HT), diabetes mellitus (DM), hyperlipidaemia, and smoking; genetic predisposition; compliance with medications; and comorbidities were obtained at the time of angiography. Laboratory data were recorded on pre-prepared study forms. Glucose (mg/dL), triglyceride (mg/dL), total cholesterol (mg/dL), low-density lipoprotein cholesterol (mg/dL), high-density lipoprotein cholesterol (HDL-C) (mg/dL), and haemoglobin (g) levels were obtained at the time of angiography. Patients were considered hypertensive if they had a previous diagnosis of HT or used any antihypertensive medications. A prior history of DM or treatment with antidiabetic medication was used to confirm the diagnosis of DM. The study was approved by the ethics committee of the University of Health Sciences, Dr.Siyami Ersek Thoracic, and Cardiovascular Surgery Training and Research Hospital (Number: E-28001928-604.01.01-207661794-2023), and informed consent was obtained from all patients or the family member.

Coronary Angiography
The digital angiographic system used for performing coronary angiography was the AXIOM Sensis (Siemens AG, Munich, Germany). The decision to perform coronary angiography was made based on the results of non-invasive stress tests or a high clinical suspicion for coronary artery disease (CAD). All patients underwent selective coronary angiography using the standard Judkins technique with JL4 and JR4 6 French catheters. The access site was the femoral artery in all cases (Ten years ago, all cases in our hospital were performed through femoral access for economic reasons). All patients received Iopromide (Ultravist 370; Schering AG, Berlin, Germany) as the preferred contrast agent. The average injection volume was 5-9 mL of opaque material. The coronary arteries were visualized at a rate of 15 frames/second-from both right and left oblique positions-using cranial and caudal angles.

CSF and Thrombolysis in Myocardial Infarction (TIMI) Frame Count (TFC)
We defined CSF based on the criteria described by Beltram, including (a) absence of obstructive epicardial CAD, (b) TFC > 27 frames, and (c) delayed distal vessel contrast opacification of epicardial coronary arteries [3].
Two independent and experienced interventional cardiologists, who were blinded to the study, reviewed all angiograms, and the TFC was calculated for each patient. TFC was first defined by Gibson and refers to the number of cine frames required for the contrast material to reach the distal landmarks. These distal landmarks included the distal bifurcation of the LAD (known as the "moustache"), the distal bifurcation of the longest lateral left ventricular wall artery branch for the circumflex artery (Cx) (Figure 1), and the first posterolateral artery branch of the right coronary artery (RCA) ( Figure 2) [8].
angiography using the standard Judkins technique with JL4 and JR4 6 French catheters. The access site was the femoral artery in all cases (Ten years ago, all cases in our hospital were performed through femoral access for economic reasons).All patients received Iopromide (Ultravist 370; Schering AG, Berlin, Germany) as the preferred contrast agent. The average injection volume was 5-9 mL of opaque material. The coronary arteries were visualized at a rate of 15 frames/second-from both right and left oblique positions-using cranial and caudal angles.

CSF and Thrombolysis in Myocardial Infarction (TIMI) Frame Count (TFC)
We defined CSF based on the criteria described by Beltram, including (a) absence of obstructive epicardial CAD, (b) TFC > 27 frames, and (c) delayed distal vessel contrast opacification of epicardial coronary arteries [3].
Two independent and experienced interventional cardiologists, who were blinded to the study, reviewed all angiograms, and the TFC was calculated for each patient. TFC was first defined by Gibson and refers to the number of cine frames required for the contrast material to reach the distal landmarks. These distal landmarks included the distal bifurcation of the LAD (known as the "moustache"), the distal bifurcation of the longest lateral left ventricular wall artery branch for the circumflex artery (Cx) (Figure 1), and the first posterolateral artery branch of the right coronary artery (RCA) ( Figure 2) [8].  CSF was defined as TFC > 2 standard deviations from the published normal range. Given that the LAD is longer than the other coronary arteries, the TFC of the LAD was divided by 1.7 and a corrected TFC (CTFC) was calculated [8]. The CTFC values were utilized in further analyses. The mean TFC was calculated by taking the average of the LAD, Cx, and RCA TFCs.
Right anterior oblique projections with caudal angulation (RAO caudal view) were used for the LAD and Cx. A left anterior oblique projection with cranial angulation (LAO cranial view) was used for the RCA. The normal TFC values were 36.28 ± 2.6 frames for the LAD, 22.28 ± 4.1 frames for the Cx, and 20.48 ± 3 frames for the RCA [8]. These values were described when cineangiography was performed at an acquisition rate of 30 frames/second. Given that the acquisition rate in our study was 15 frames/second, we multiplied the TCF values by 2 to adjust our values to the 30 frames/second values [8]. CSF was defined as TFC > 2 standard deviations from the published normal range. Given that the LAD is longer than the other coronary arteries, the TFC of the LAD was divided by 1.7 and a corrected TFC (CTFC) was calculated [8]. The CTFC values were utilized in further analyses. The mean TFC was calculated by taking the average of the LAD, Cx, and RCA TFCs.
Right anterior oblique projections with caudal angulation (RAO caudal view) were used for the LAD and Cx. A left anterior oblique projection with cranial angulation (LAO cranial view) was used for the RCA. The normal TFC values were 36.28 ± 2.6 frames for the LAD, 22.28 ± 4.1 frames for the Cx, and 20.48 ± 3 frames for the RCA [8]. These values were described when cineangiography was performed at an acquisition rate of 30 frames/second. Given that the acquisition rate in our study was 15 frames/second, we multiplied the TCF values by 2 to adjust our values to the 30 frames/second values [8].

Statistical Methods
The study population's general characteristics were obtained by conducting descriptive analyses. The Shapiro-Wilk test was used to determine the distribution of the variables; a normal distribution was observed for all variables. Therefore, independent two-sample t-tests were conducted to compare continuous variables with two groups. Continuous variables were compared using one-way analysis of variance (ANOVA) with Scheffe's post-hoc test. A multiple logistic regression model was used to identify covariates associated with cardiovascular (CV) and non-CV death in patients with CSF. The mean ± standard deviation was used to present continuous variables. The Chi-squared test was utilized to compare categorical variables. Counts and percentages were used to present categorical variables. p-values < 0.05 were considered significant. Analyses were performed with SPSS software, version 23.0 (IBM Corp.; Armonk, NY, USA).

Statistical Methods
The study population's general characteristics were obtained by conducting descriptive analyses. The Shapiro-Wilk test was used to determine the distribution of the variables; a normal distribution was observed for all variables. Therefore, independent two-sample t-tests were conducted to compare continuous variables with two groups. Continuous variables were compared using one-way analysis of variance (ANOVA) with Scheffe's posthoc test. A multiple logistic regression model was used to identify covariates associated with cardiovascular (CV) and non-CV death in patients with CSF. The mean ± standard deviation was used to present continuous variables. The Chi-squared test was utilized to compare categorical variables. Counts and percentages were used to present categorical variables. p-values < 0.05 were considered significant. Analyses were performed with SPSS software, version 23.0 (IBM Corp.; Armonk, NY, USA).
Significant differences in LAD (p < 0.001), Cx (p = 0.002), RCA, and mean TFC in patients with CV vs. non-CV causes of death were observed ( Table 2). Multiple logistic regression analyses were conducted to determine the independent predictors of long-term cardiovascular (CV) mortality and total mortality among patients with CSF. Total mortality was associated with age (odds ratio [  In the multiple regression model of patients with CSF, mean TFC was also associated with CV mortality (OR: 3.318; 95% CI: 1.101-10.000) (Figure 3). Specifically, the risk of CV-related mortality increased 3.3-fold with every 1-frame increase in mean TFC (Table 4).

Discussion
Our study revealed that patients with CSF have a high prevalence of comorbidities, CV risk factors, and CV-related mortality after 10 years of follow-up. To the best of our knowledge, this study is the first to examine long-term mortality in patients with CSF. In patients with CSF, 7% experienced CV-related mortality, whilethe all-cause mortality rate was 15.3%. Recurrent myocardial infarction, ACS, hospitalizations, and coronary inter-

Discussion
Our study revealed that patients with CSF have a high prevalence of comorbidities, CV risk factors, and CV-related mortality after 10 years of follow-up. To the best of our knowledge, this study is the first to examine long-term mortality in patients with CSF. In patients with CSF, 7% experienced CV-related mortality, whilethe all-cause mortality rate was 15.3%. Recurrent myocardial infarction, ACS, hospitalizations, and coronary interventions were frequently observed. The determinants of total mortality were age, HT, DM, discontinuation of medications, and low HDL-C levels. Notably, medication noncompliance was the most important factor determining CV-related mortality in patients with CSF. Other significant determinants of CV mortality included low HDL-C levels and age.
Patients with CSF frequently presented with ischemia and recurrent chest pain. The pathogenesis of CSF remains to be elucidated; however, several studies suggest that atherosclerosis, inflammation, and endothelial dysfunction are underlying mechanisms of CSF [7].
The current study addresses the limitations of current luminology practices. CSF, atherosclerosis, and CAD exhibit common CV risk factors and clinical presentations. Therefore, increased long-term mortality in patients with CSF is expected [2]. CSF is a common finding on diagnostic coronary angiography (1-7% of all coronary angiographies) [2]. In patients who underwent CAG for chest pain, CSF is a simple, valuable, and additional biomarker that comes at a low cost [9]. In our study, eight patients required implantation of coronary stents for severe CAD and two patients required a CABG procedure during the long-term follow-up period. CSF is reportedly a risk factor for sudden cardiac death, arrhythmia, myocardial ischemia, ST elevation, and myocardial infarction [7,10,11]. CSF is also more common in males, cigarette smokers, and patients with increased CV risk [3]; however, the long-term mortality rate of patients with CSF is still unknown.
While patients with CSFchest pain and normal arteries are generally considered to have a favourable prognosis, some studies have suggested that these patients may be vulnerable to ACS [3,4,12]. Specifically, in some study cohorts, approximately two-thirds of patients with CSF presented with ACS [2]. In our study, we observed that 27 patients (19.7%) were readmitted to the hospital due to ACS. Our observations indicate that CSF is associated with recurrent coronary events and poor survival, particularly in patients who are non-compliant with medications. Atak et al. demonstrated that patients with CSF have an increased rate of QT dispersion, which is associated with the risk of ventricular arrhythmias and CV mortality [13]. In addition, sudden cardiac death due to ventricular arrhythmias has been reported in patients with CSF [6,13]. These findings are consistent with our results, as one patient died from sudden cardiac death in our study.
The definite mechanism of CSF is still uncertain; however, endothelial dysfunction and increased microvascular resistance may play a role. Previous studies have reported significant differences in fractional flow reserve values and pressure between the proximal and distal coronary arteries, indicating higher resistance in the micro-circulation [14]. In addition, Fineschiet al. used invasive hemodynamic measurements to show that resting microvascular resistance is elevated in patients with CSF [15]. Beltrame et al. found evidence that CSF is linked to a sustained increase in resting coronary microvascular tone [16]. Several studies have also demonstrated that patients with CSF exhibit impaired endothelium-dependent flow-mediated dilatation, suggesting that coronary vascular endothelial dysfunction could be an important mechanism underlying CSF pathogenesis [14,17]. The presence of microvascular disease in patients with CSF has been demonstrated through myocardial biopsy studies [18]. In these studies, endothelial thickening and lumen stenosis were observed. Therefore, CSF may represent an early, subclinical form of atherosclerosis [7]. Indeed, two studies utilizing advanced imaging techniques, such as intravascular ultrasound, have demonstrated that CSF may indicate the presence of diffuse, non-obstructive atherosclerotic disease in the coronary arteries [19,20]. The findings from the present study support this hypothesis; over time, indistinct atherosclerotic plaques can become obstructive atheroscle-rotic plaques and lead to vascular occlusion, particularly in patients who are non-compliant with their medications.
In this study, we observed that patients who discontinued their medications hadhigher mortality rates. Specifically, the most important factor determining CV-related causes of death in patients with CSF was discontinuation of medications. These findings demonstrate that the treatment and follow-up of patients with CSFshould be equivalent to that of patients with CAD. In this study, nearly 40% of patients haddiscontinuedtheir medications during follow-up, likely contributing to the etiology of recurrent coronary events and poor survival. Previous studies have shown that non-compliance with cardioprotective medications (e.g., ACEI, statins, and beta-blockers) among outpatients with CAD was associated with elevated risk of mortality from all causes [21,22]. Compliance with medical treatment plays a crucial role in achieving favorable clinical outcomes in patients with CAD, such as a decreased re-admission rate and reduced CV-related morbidity and mortality [23].
Another determinant of CV and all-cause mortality in this study was low HDL-C levels. In the literature, an association between HDL-C levels and longevity has been proposed [24]. Likewise, other studies have revealed that high HDL-C levels prevent many age-related diseases, whereas reduced levels of HDL-C are associated with an increased risk of coronary heart disease [25]. Indeed, epidemiologic studies have revealed a negative linear correlation between HDL-C levels and mortality [26]. Moreover, in a meta-analysis, Gordon et al. demonstrated that for every 1 mg/dl (0.026 mmol/L) increase in plasma HDL-C levels, the risks of CHD and CV-related mortality were reduced by 2-3% and 3.7-4.7%, respectively [26]. The Framingham Study also revealed an inverse correlation between HDL-C levels and both CV disease and total mortality [27]. The anti-inflammatory, antioxidant, anti-apoptotic, anti-thrombotic, and vasodilatory properties of HDL-C may explain its atheroprotective characteristics [28,29]. In accordance with the literature, HDL-C levels were correlated with CV-related mortality and total mortality in the present study. Consistent with our results, Hawkins et al. demonstrated that patients diagnosed with CSF had significantly lower HDL-C levels [30]. In light of these findings, it is reasonable to conclude that patients with CSF should be encouraged to make lifestyle modifications, such as smoking cessation, participation in regular physical activity, and consumption of a healthy diet-in addition to receiving medical treatment.
Several clinical observational studies demonstrate that hyperglycaemia is associated with poor outcomes in patients with obstructive CAD [31,32]. On the other hand, patients with angiographically normal coronary arteries with slow flow represent a population with a potentially different etiology and pathophysiology. Recent studies define MINOCA (Myocardial Infarction with Non-Obstructive Coronary Arteries) as a separate clinical entity [33][34][35]. MINOCA and Ischemia with No Obstructive Coronary Artery Disease (INOCA) are commonly observed findings after coronary angiography [36].
The association between poor clinical outcomes and high blood glucose appears to be different in obstructive versus non-obstructive CAD. In patients who present with myocardial infarction and obstructive coronary arteries (MIOCA), hyperglycaemia is associated with poor outcomes [37]; however, the prognostic impact of hyperglycaemia in INOCA/MINOCA remains controversial.
In our study, in patients with normal coronary arteries, the history of DM did not affect the presence or the severity of CSF. Further studies are required to explore the differences in the pathophysiology of MINOCA, INOCA, and MIOCA. Both INOCA and CSF are heterogeneous clinical entities. Multiple risk factors such as inflammation and thrombosis are likely to interplay to determine the poor outcomes.Similar to our findings, MINOCA patients display poor outcomes with a substantial risk of recurrent major adverse cardiac events (MACE) during follow-up [38]. In patients with MINOCA, there is evidence of pro-thrombotic activity and coronary microvascular dysfunction [39].
Our study adds to the knowledge about CSF outcomes in the literature in patients with CSF-in that these patients experienced poor survival. The limitations of our study include a small sample size and differences in the prevalence of risk factors such as age, HT, and DM among groups. Moreover, since we could not reach all recorded patients, we could not create an age-and sex-matched control group. CSF is a very complex phenotype that extends beyond known CAD risk factors such as thrombosis and inflammation. Novel indices and biomarkers are required to improve our understanding of CSF. Furthermore, larger randomized controlled studies are required to confirm our observations.

Conclusions
Patients with CSF experienced high CV-related mortality and total mortality after 10 years of follow-up. Age, HT, DM, discontinuation of medications, and low HDL-C levels were risk factors for mortality in CSF patients. Our findings support the notion that coronary angiography findings extend beyond luminology. Further studies are required to confirm CSF as a therapeutic target in cardiovascular disease.