Coronary Plaque Characteristics and Cut-Off Stenosis for Developing Spasm in Patients with Vasospastic Angina

Coronary plaque distribution, frequency and cut-off value of percent stenosis for developing vasospasm are uncertain in patients with vasospastic angina (VA). We enrolled 2960 patients who received coronary angiography (CAG) and ergonovine provocation test prospectively in 11 university hospitals in Korea. A total of 1836 patients with VA and 867 without VA were included. Plaque and % stenosis were defined as ≥1% luminal narrowing and mean of each segmental stenosis. Overall frequency of plaque and % diameter stenosis was compared among VA-patients with index coronary spasm positive, those with index arterial spasm negative/other arterial spasm positive (INOP) and non-VA patients. Diameter stenosis associated with the spasm positivity was investigated. Overall plaque frequency and % stenosis were higher in VA patients than non-VA patients. Plaque frequency was 27.6% (243/881) in spasm positive at LAD, 16.4% (157/955) in LAD INOP and 12.6% (109/867) in non-VA with statistic difference (P < 0.001). Same trend for higher rate was observed in LCx and RCA. For % stenosis, 36.6 vs 32.4% (p = 0.010) in LAD, 36.1 vs. 28% (p < 0.001) in LCx and 35.3 vs.30.0% (p = 0.047) in RCA, respectively. Diameter stenosis of LAD with spasm positive vs. LAD INOP vs. non-VA were 38.3%, 34.0%, 32% (P = 0.002) with similar pattern in LCx and RCA. By multivariate logistic regression analysis, coronary stenosis of LAD ≥ 35% or LCx ≥35% or RCA ≥ 40% were independent predictor of developing spasm (OR 2.019, 95% CI 1.315–3.100, P = 0.001). In conclusions, spastic coronary artery had more plaque frequency, higher % stenosis than in non-spastic coronary in VA patients. The spasm related and unrelated coronary in VA patients had more plaque than in matched and unmatched coronary arteries in non-VA patients. Coronary stenosis ≥35% in LAD and LCx was an independent predictor of developing spasm.


Results
A total of 1836 patients with VA and 867 with non-VA are included for overall comparison for their atherosclerosis status (Fig. 1). Baseline characteristics between VA and non-VA patients were comparable except for male sex, active smoker, percutaneous coronary intervention (PCI) history, high density lipoprotein cholesterol (HDL-C), high-sensitivity C-reactive protein (hsCRP), Troponin (Table 1). After 73 patients who do not have specific spasm location data with coronary diameter <2.5 mm were excluded in specific coronary focusing comparison, 1763 patients was included in spasm location oriented comparison with INOP and non-VA patients (Fig. 1).   (Table 3 and Fig. 2) The detailed segmental distribution of plaque was presented in Fig. 3. Segmental distribution of atherosclerotic plaque was more prevalent in spastic patients. (Fig. 3) Atherosclerosis analysis focusing on individual coronary: spasm positive at index coronary artery vs. index coronary spasm negative but positive on other coronary (INOP) vs. Non-VA patients. Diameter % stenosis is higher in VA patients than in non-VA patients at each coronary artery, 38.27  (Fig. 3).
VA-patients was always more likely to have more atherosclerosis at spastic coronary than the other coronary within same VA-patients (Tables 4, 5 and Supplementary Table 1 Table 1).
Even the spasm irrelevant coronary in VA patients are more likely to have higher plaque status parameters than those matching coronary with non-VA patients. For example, in LAD spasm positive case, LCx plaque frequency was highest in LAD spasm case (8.1%) than that of INOP (9.9%) and non-VA (3.2%) (Supplementary Table 1). Same trend was observed in plaque stenosis and burden (Supplementary Table 1).

Discussions
We investigated for the first time the detailed distribution, severity of atherosclerotic plaque and their relationship in VA patient as compared to non-VA patients. We found that the VA patients who has spasm at any coronary artery were more likely to have overall atherosclerotic plaque in all coronary arteries even in non-spastic artery as compared to non-VA patients. Spastic coronary has more likely to have plaque, higher plaque % stenosis and larger plaque burden than other coronary artery irrelevant to spasm in same VA patient. The plaque parameters were higher in VA patient both in spastic index artery and other coronary artery irrelevant to spasm than those with non-VA patients. These results are in line with those from previous small sized intravascular ultrasound (IVUS) study 6 which is only one literature regarding this kind of subject.  www.nature.com/scientificreports www.nature.com/scientificreports/ We analyzed the plaque with more focus on specific spasm positive coronary artery to compare in-depth with other non-spasm responsible coronary of VA patient as well as coronaries of non-VA patients. Spasm positive coronary had more chance to have plaque, higher diameter stenosis and higher plaque burden as compared to matching coronary artery of non-VA patients. Interesting finding is that, in our combined comparing with those with index coronary not spastic but other site positive (INOP), the likelihood of having plaque, % stenosis and atheroma burden was highest in index arterial spasm positive, middle in INOP and lowest in non-VA patients. Same trend was observed in all 3 coronary arteries. Even the plaque in INOP showed higher or higher trend than that of non-VA patients consistently.
We also, for the first time, investigated the cut-off point of coronary stenosis which might be related to spasm provocation as 35-40% and found this value was independent predictor for spasm.
Our study results imply that the spasm is very closely linked with plaque. Spasm status might be associated with the whole vascular atherosclerotic circumstance even in non-spasm related artery as well as spasm culprit artery. This could reflect the systemic nature of vasospasm which is linked with atherosclerosis.
The concept that plaque can be a trigger for vasospasm or reversely, spasm itself can initiate or aggravate the atherosclerosis is already well known by experimental and clinical studies 7 . The spasm itself could cause atherosclerotic progression, plaque injury or rupture and could even lead to ST-elevation myocardial infarction from clinical or experimental evidences 1,[8][9][10][11] . It is postulated that the mechanical insult by spasm on the coronary might have some role. Another similar mechanism is myocardial bridging which can cause a vascular event by repetitive mechanical insult to coronary artery. This coincided well with the hypothesis that vasospasm could induce mechanical injury of coronary artery even resulting in acute thrombotic events 12 . www.nature.com/scientificreports www.nature.com/scientificreports/ On the contrary, atherosclerotic plaque can induce and aggravate vasospasm mainly driven by endothelial dysfunction, nitric oxide (NO) deficiency and mild inflammation of atherosclerosis and vascular smooth muscle cell (VSM) hyper-reactivity 13 . One IVUS-virtual histology study found that coronary segments with endothelial dysfunction had larger necrotic core plaques and the size of necrotic core plaques was the principal determinant of coronary endothelial dysfunction 14 . However the causal relationship of which one between spasm and atheroma is initiator has been uncertain.
Recent study arguing perivascular fat has a role in spasm also support the atherosclerosis could influence the spasm by VSM contraction pathway 15 . But we consider more important factors for spasm is intra-lumen plaque which could more directly affect the luminal circumstances as resulting in flow limitation and obstruction. Moreover peri-vascular adipose tissue could just be a reflection of vascular intimal atherosclerosis 16 . Thus, we think, intramural and intraluminal atherosclerosis might be more critical as our study results. We can also postulate that in view of the fact that the plaque burden is prevalent and higher in VA patients even in the spasm unrelated artery than that of non-VA patients, the atherosclerosis might have initiate and precipitate the vasospasm. In other words, the atherogenic milieu first and later on, the spasm occurred at site of minimal stenosis around 35-40% of lumen narrowing of coronary diameter, with moderate endothelial dysfunction and combined VSM hyper-reactivity. This concept is partly supported by very few evidence yet, and need to be further investigated 17 .    www.nature.com/scientificreports www.nature.com/scientificreports/ A study reported that VA patients with spasm at the site of plaque have poorer clinical outcomes than those with spasm at other site 18 . In that study, the authors defined the plaque as 75% or more diameter stenosis which is usually accepted cut-off vale of traditional atherosclerotic stenosis. However this value may not be true in the vasospasm, rather less luminal narrowing would be appropriate because VA patients could experience more severe stenosis enough to make flow limitation with vascular contraction on top of mild fixed lesion 19 . A research indicating even non-obstructive coronary artery disease (CAD) among non-VA population with angiographic stenosis 20-70% has nearly 3-fold increased risk of death compared with those without apparent CAD also support our hypothesis 20 . Thus less than 75% of spasm related atherosclerotic stenosis might be appropriate in spasm occurrence and determining prognosis. With this curiosity, we firstly aim to pursue the true cut-off point of diameter stenosis of spasm provocation in VA suspicion patients, and we are preparing for the cut-off diameter stenosis for determining poor clinical outcome in VA-patients.
In another view point of the plaque burden, we could postulate that much plaque burden could not provoke the spasm because the hyperactive vascular response hardly occurred due to lack of remaining endothelial dysfunction to exert spasm and few residual functional vascular smooth muscle cell in arterial media nor the vasoconstriction could not overcome the coronary arterial stiffness and volume of the plaque. Thus, we hypothesized   www.nature.com/scientificreports www.nature.com/scientificreports/ that diameter stenosis which is related to development of vasospasm would be around 30-70% arbitrarily. With this assumption we tested the spasm prone cut-off % stenosis in each coronary artery and found appropriate value of ≥35% both in LAD and LCx and ≥40% in RCA which were remained meaningful by multivariate logistic regression analysis despite of little sensitivity. We assume that with this cut-off value, both VSM dysfunction and endothelial dysfunction start exerting effect on the vascular spasm.
Clinical implication. The disease entity of VA is still remained to be elucidated. Some researchers suggested genetic background is closely associated with VA occurrence. For example, a person with lack of activity of aldehyde dehydrogenase (ALDH) 2 owing to having variant ALDH2*2 genotype which are related with alcoholic flushing syndrome are prone to suffer from VA 21 . Among lots of plausible mechanism and explanation for developing VA, we concentrate on the relationship between atherosclerosis and VA.
Our results could imply that the atherosclerotic plaque occupying around 1/3 of lumen could predict the spasm positive in VA suspicion patients. This could be of value in clinical practice for our assessing and predicting spasm positivity in ergonovine (EG) provocation test especially in case of the EG provocation response is obscure. We could also be more alert in those patients and may pay more attention to than those who have plaque less than 30-40% especially in case of uncertain EG provocation test results. In treatment view point, the atherosclerosis involved spasm positive patients ought to be candidate for statin therapy despite of the uncertainty of statin's role in VA, because there is still possibility of minimal atherosclerotic plaque could have contribute spasm initiation or intensification from indirect data from our study. The role of statin in patients with plaque narrowing ≥ 35-40% of lumen ought to be further investigated in future study. Further research is on-going as to whether this value could predict prognosis of VA patients by our study group.

Study Strength.
Our study has some strength.
Firstly, we used the provocation methods with EG other than acetylcholine (ACH). Although both agents are useful and widely utilized in clinical practice, the EG might be better because it provokes vasospasm via VSM contraction not by endothelial mediated vasoconstriction which could be induced by ACH. The endothelial dysfunction could be usually revealed by ACH provocation test at early stage of atherosclerosis 22 , so the spasm could not occur in moderate to severe atherosclerotic artery. Moreover, some study argues that the endothelial dysfunction is not main mechanism, instead VSM hyper-constriction by RHO kinase activation is a major player 23,24 .
Secondly, we firstly evaluated the association of atheroma in coronary artery with spasm in detail using large scale prospective multicenter data and found the cut-off values of spasm inducing atheroma burden in VA patients with atheroma raging 35-40% which is lower what has been considered to be significant in general atherosclerosis patients.
Thirdly, we assessed the plaque in detail with 3 aspects like presence of plaque, dimeter % stenosis of coronary lumen and plaque burden expressed as summation of each segmental % stenosis and showed consistent results in all aspect of comparisons.
Limitations. Firstly, our study was performed using registry data which has inevitable biases. For example, in our registry, patients who complain of chest pain suspicious of VA received coronary angiography and ergonovine provocation test at the physician's discretion. Thus in some patients, the final diagnosis was not VA. Accordingly, selection bias can exist because the non-VA patients might have other hidden morbidity that can provoke chest pain other than coronary spasm, i.e. they could not have definitely normal coronary artery. However our registry is large scale, prospective and multicenter one that has value.
Secondly, the percent stenosis and burden of plaque was just estimated by visual and QCA not by IVUS or OCT which are more correct. Thus the measured value was expressed with unit of number 5. We also do not measure correct plaque volume, instead we employed the indirect methods of diameter stenosis and segmental stenosis summation. Nevertheless we perform QCA by independent person who are blind to the study performance in central laboratory facilities.
Thirdly, the cut-off value for sensitivity and specificity of lumen stenosis is low. Fourthly, the rate of patients with hypertension, diabetes and chronic kidney disease are lower than other study. We consider this could be attributed to selection bias, recall bias or different population enrolled with only suspicious of VA which might have a little risk factors than those in other clinical studies.

Methods
Study objective. We aim to investigate the nature of atherosclerosis in VA patients like plaque frequency, distribution, burden and compare these with those of non-VA patients and index coronary arterial negative but other site coronary positive VA patients in detail. The final objective of the study is to find the insight of association or causal relationship between spasm and atherosclerosis in VA patients.
Study protocol was approved by ethics committees/institutional review board of each participating hospital including Hallym University Sacred Heart Hospital Institutional Review Board, named 'HUMC ethics committees & IRB' (IRB No. 2010-I007) and all patients gave written informed consent. All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.
Patients. VA-Korea (Variant angina Korea) registry is a nation-wide prospective multicenter registry which enrolled patients with chest pain suspicious of VA who received coronary angiography (CAG) and EG provocation test 25,26 . Suspicion of VA was solely at a physician's discretion. Hence patents with normal or atherosclerotic Scientific RepoRtS | (2020) 10:5707 | https://doi.org/10.1038/s41598-020-62670-z www.nature.com/scientificreports www.nature.com/scientificreports/ coronary artery can be included in this registry other than those with VA. Adults aged 18 or over were candidate for enrollment. Patients having malignancy, end stage renal disease on dialysis, inflammatory disease and catheter-induce spasm at baseline CAG were excluded. Hypertension was defined as either blood pressure ≥140/90 mmHg or the current use of anti-hypertensive medications. Diabetes was defined as having fasting glucose ≥126 mg/dL or hemoglobin A1 C ≥o6.5% or current intake of anti-diabetes medications. Dyslipidemia was defined as either total cholesterol ≥240 mg/dL or the current use of anti-dyslipidemia drugs.
Study protocol was approved by institutional review board of each participating hospital and all patients gave written informed consent. A total of 2960 patients were registered consecutively from May 2010 to June 2015 in 11 tertiary hospitals in Korea with high volume CAG and percutaneous coronary intervention (PCI) performances. Among them, 257 had missing data and 2703 were included in analysis. A total of 1836 patients were diagnosed to have positive results in their provocation tests 867 had negative results (Fig. 1). Each spastic lesion, 881 in LAD, 419 in LCx, 942 in RCA were included in lesion specific analysis (Fig. 1).
Coronary angiography and EG provocation test. We followed provocation methods from Japanese Circulation Society guideline for diagnosis and treatment of patients with VA 27 . Detailed methods of CAG, EG provocation test and adjudication of EG provocation test results are reported previously 25,26 .
Quantitative coronary angiography (QCA) was performed twice, firstly in each responsible hospital by the personnel who are not aware of the study and secondly by the dedicated physicians who were blinded to the study at core laboratory located in Seoul St. Mary's Hospital using dedicated QCA program and by visual assessment.
Coronary plaque analysis. Firstly, we compared the overall plaque status of coronary arteries like distribution, frequency and % diameter stenosis between VA patients and non-VA patients. Baseline characteristics between VA and non-VA patients were compared (Table 1).
Secondly, we focused on each spastic coronary artery. We compared plaque status (including plaque frequency, mean diameter % stenosis, total plaque burden) between spastic artery and non-spastic artery in same VA-patient. We also compared spasm positive arteries of VA-patients with index arteries which are negative spasm but other site spasm positive one [INOP] or non-VA patients regarding to plaque status. We also compared spasm related index artery and non-spasm related artery with corresponding arteries of non-VA patient.
In this spasm related coronary focused analysis, we only included patients with coronary diameter of ≥2.5 mm artery which had available data on spasm location (location mean name of coronary like LAD, LCx, RCA). Patients having multi-vessel spasm are included several times for each arterial analysis (Fig. 1). For example, if a patient had both LAD and LCx spasm, their atherosclerosis on LAD and LCx were both analyzed as individual coronary artery. Thus the LAD spasm 881, LCx 419, RCA 942 lesions were included in analysis. Overlapped spasm positive coronary arteries ranging from 2 to 4 times were described in detail in Fig. 1 study flow. Thirdly, we are also pursue to find the atherosclerotic % stenosis which could induce vasospasm with sensitivity and specificity test starting form 30% until 70% step by step with 5% increment of diameter stenosis in each coronary artery by comparing diameter % stenosis between index arterial spasm positive patients and index arterial spasm negative patients (which included INOP and non-VA patients). We also tested this cut-off value of each coronary artery and composite cut-off values of LAD or LCx or RCA by multivariate logistic regression analysis.
Coronary percent diameter stenosis was estimated by visual estimation and QCA in coronary angiography in central lab. Plaque was defined as ≥1% luminal narrowing of coronary diameter. If a coronary artery had at least one plaque in any segment of the same coronary artery, that was defined as having a plaque. If at least one plaque was present among 3 coronary arteries, a patient was regarded to have plaque. Mean % diameter stenosis of each coronary artery was calculated by averaging each % stenosis of proximal, medial and distal segment. Plaque burden was defined as summation of % stenosis of each segment, i.e. the maximum 300 and minimum 1. We do not have intravascular ultrasound (IVUS) data and could not exactly estimate the plaque volume, thus we adopt this indirect methods of plaque summation. Statistical analysis. Continuous variables are expressed as mean ± standard deviation and categorical variables are presented as frequencies and percentages. For between-group comparisons, independent t-test or Mann-Whitney U-test were used for continuous variables, and the chi-square test or Fisher's exact test were used for categorical variables. A simple logistic regression and multivariate logistic regression model were used to identify independent predictors of cut-off values of diameter percent stenosis for determining vasospasm.
We performed backward stepwise logistic regression analysis using variables in Table 1 in testing the each variable of LAD ≥ 35%, LCx ≥ 35%, and RCA ≥ 40% as independent risk factors for inducing spasm respectively. We also conducted logistic regression analysis with same manner in testing composite of LAD ≥ 35% or LCx ≥ 35% or RCA ≥ 40% as independent risk of provoking spasm. Gender, age, diabetes and LAD ≥ 35% were significant in testing LAD ≥ 35%. Gender, dyslipidemia and LCx ≥ 35% were also significant in LCx ≥ 35%. Gender, age, diabetes and smoking were found to be significant in testing RCA ≥ 40%. The above variables and the variables that investigator considered being important such as age, hypertension, diabetes, smoking, history of using calcium channel blockers were included in the multiple logistic regression analysis as fixed variables. Statistical analysis was carried out by using SPSS ver. 23 (IBM Corporation).
conclusions VA patient has more probability of having atherosclerotic plaque, higher lumen stenosis and larger plaque burden even in non-spasm related coronary artery than in non-VA patients. Atherosclerotic lumen narrowing of ≥35-40% in coronary artery could be a cut-off value for development coronary vasospasm.