case study about coronary artery disease

The patient in this case exhibits poor glycemic control with a HbA 1c above target (<7%), which has additional implications for his cardiovascular risk. Among patients with type 2 diabetes, the relative risk for fatal coronary heart disease is 1.16 for each one-percentage point increase in HbA 1c . 4 Additionally, his age, gender, elevated total cholesterol, low HDL-C, and hypertension further increase his risk of significant cardiovascular disease. 5 This patient, like many patients with diabetes and insulin resistance, has a lipid profile characterized by low HDL-C, high triglycerides, and a modest increase in LDL-C, which are predominantly small dense particles. The overall particle number is increased as would be reflected in an elevated apolipoprotein B (apoB) level. These small, dense LDL particles are more easily oxidized, enter the arterial wall more readily, are cleared more slowly, and are thus considered more atherogenic. His physical exam is suggestive of peripheral arterial disease (PAD), and coincident coronary artery disease is found in up to 90% of patients with definitive PAD. 6 Aggressive treatment of these cardiovascular risk factors can improve both morbidity and mortality. In the Steno-2 trial, 160 patients with type 2 diabetes were randomized to either intensive therapy of conventional therapy. 7 Targets for the intensive therapy group included HbA 1c <6.5%, fasting total cholesterol <175 mg/dL, fasting triglycerides <150 mg/dL, systolic blood pressure <130 mm Hg, and diastolic blood pressure <80 mm Hg. After a mean of 13.3 years of follow up, intensive therapy was associated with a decrease risk of cardiovascular death (hazard ratio 0.43) and cardiovascular events (hazard ratio 0.41). Therefore, our patient warrants intensive lipid, blood pressure, and glycemic control independent of his angina and stress test results.

Our patient's testing is suggestive of flow-limiting coronary artery disease (CAD) in what is likely a right coronary artery distribution. His symptoms are consistent with stable angina rather than an acute coronary syndrome. Several studies have examined the utility of revascularization in addition to optimal medical therapy (OMT) in patients with stable angina. In A Very Early Rehabilitation Trial (AVERT), patients with CAD (defined as 50% or greater stenosis of one or more coronary arteries) and an LDL of 115 mg/dL or greater were randomized to either percutaneous coronary intervention (PCI) or aggressive lipid lowering with high-dose atorvastatin. 8 The study population was 84% male and 95% Caucasian, and 78% of those enrolled experienced angina at baseline. After 18 months of follow up, the atorvastatin group exhibited a 36% decrease in ischemic events. The Bypass Angioplasty Revascularization Investigation 2 Diabetes (BARI 2D) study group investigated the role of PCI in addition to OMT in patients with type 2 diabetes and stable ischemic heart disease. Patient with left main disease were excluded from the trial. In this study population, PCI provided no improvement in five-year survival or the rate of major cardiovascular events. 9 Similarly, the Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation (COURAGE) trial found that PCI in addition to OMT did not significantly decrease the rate of major cardiovascular events when compared to medical therapy alone in patients with stable CAD. 10 Again patients with severe left main stenosis were excluded from the study. Per the 2012 ACC/AHA guidelines on stable ischemic heart disease, diabetic patients with stable angina should receive medical therapy, including aspirin, beta-blocker, moderate-to-high-intensity statin, and either a long-acting nitrate or calcium channel blocker. 11 Furthermore, our patient's stress test is highly suggestive of single vessel disease without left main or proximal left anterior descending (LAD) artery involvement. Therefore, at this point in the clinical scenario, option A is incorrect, and option D is the next most appropriate step.

If symptoms persist despite medical therapy, it is reasonable to pursue revascularization. In patients with three-vessel CAD or left main disease, PCI has an increased rate of major adverse cardiac and cerebrovascular events compared to CABG. 12 The Future Revascularization Evaluation in Patients with Diabetes Mellitus: Optimal Management of Multivessel Disease (FREEDOM) Trial compared revascularization with coronary artery bypass graft (CABG) versus PCI in patients with diabetes and severe CAD. In this population, the majority of whom had multivessel CAD, CABG resulted in a significant decreased rate of major adverse cardiac or cerebrovascular events after one year (12.4 vs. 17.8%). 13 Patients also report improvements in angina frequency, physical limitations, and quality of life with CABG compared to PCI. 14,15

Cardiac rehabilitation plays an important role in secondary prevention as well as ongoing risk factor modification. In fact, data show that rates of death and MI decrease with increased attendance at rehabilitation sessions. 16 Stable angina, as in this patient, is an appropriate indication for referral to cardiac rehabilitation, but he should first receive maximal medical therapy to treat his anginal symptoms prior to referral. Therefore, option B is incorrect.

Option C is incorrect. While an assessment of left ventricular function can be an important part of this patient's management, there were no signs or symptoms suggestive of congestive heart failure on exam. Therefore, an echocardiogram should not be the first step in his management.

Clinicians are increasingly using coronary computed tomography angiography (CCTA) to evaluate for the presence of CAD. In a study of 122 patients with diabetes, atherosclerotic plaques were identified in 95% of subjects using CCTA. 17 Similarly, an analysis of 138 patients with type 2 diabetes undergoing CCTA found that 75.2% exhibited coronary disease in at least two vessels with LAD artery involvement in 35.9% of the study population. 18 A meta-analysis of 9,592 patients found that obstructive CAD (>50% luminal stenosis) on CCTA had a positive likelihood ratio of 1.70 for major adverse cardiac events. 19 Conversely, the absence of obstructive CAD had a negative likelihood ratio of 0.008. In our patient, however, we have already found evidence of moderate ischemia on stress imaging. Per the 2010 American College of Cardiology/American Heart Association guidelines on CCT, subsequent CCTA after stress imaging showing this degree of ischemia is classified as an inappropriate use of the test. 20 Therefore, option E is incorrect.

• Preis, SR, Hwant, S, Coady, S, et al. Trends in all-cause and cardiovascular disease mortality among women and men with and without diabetes mellitus in the Framingham Heart Study, 1950-2005. Circulation 2009;119:1728-35.
• Haffner, SM, Lehto, S, Ronnemaa, T, et al. Mortality from coronary heart disease in subjects with type 2 diabetes and in nondiabetic subjects with and without prior myocardial infarction. N Engl J Med 1998;339:227-34.
• Creager, MA and Luscher, TF. Diabetes and vascular disease: pathophysiology, clinical consequences, and medical therapy: part I. Circulation 2003;108:1527-32.
• Selvin, E, Marinopoulos, S, Berkenblit, G, et al. Meta-analysis: glycosylated hemoglobin and cardiovascular disease in diabetes mellitus. Ann Intern Med 2004;141:421-31.
• Wilson, PWF, D'Agostino, RB, Levy, D, et al. Prediction of coronary heart disease using risk factor categories. Circulation 1998;97:1837-47.
• Golomb, BA, Dang, TT, and Criqui, MH. Peripheral arterial disease: morbidity and mortality implications. Circulation 2006;114:688-99.
• Gaede, P, Lund-Andersen, H, Parving, H, et al. Effect of a multifactorial intervention on mortality in type 2 diabetes. N Engl J Med 2008;358:580-91.
• Pitt, B, Waters, D, Brown, WV, et al. Aggressive lipid-lowering therapy compared with angioplasty in stable coronary artery disease. N Engl J Med 1999;341:70-6.
• The BARI 2D Study Group. A randomized trial of therapies for type 2 diabetes and coronary artery disease. N Engl J Med 2009;360:2503-15.
• Boden, WE, O'Rourke, RA, Teo, KK, et al. Optimal medical therapy with or without PCI for stable coronary disease. N Engl J Med 2007;356:1503-16.
• Fihn, SD, Gardin, JM, Abrams, J, et al. 2012 ACCF/AHA/ACP/AATS/PCNA/SCAI/STS guideline for the diagnosis and management of patients with stable ischemic heart disease. J Am Coll Cardiol 2012;60(24):e44-e164.
• Serruys, PW, Morice, M, Kappetein, AP, et al. Percutaneous coronary intervention versus coronary-artery bypass grafting for severe coronary artery disease. N Engl J Med 2009;360:961-72.
• Farkouh, ME, Domanski, M, Sleeper, LA, et al. Strategies for multivessel revascularization in patients with diabetes. N Engl J Med 2012;367:2375-84.
• Abdallah, MS, Wang, K, Magnuson, EA, et al. Quality of life after PCI vs CABG among patients with diabetes and multivessel coronary artery disease: a randomized clinical trial. JAMA 2013;310,15:1581-90.
• Cohen, DJ, Van Hout, B, Serruys, PW, et al. Quality of life after PCI with drug-eluting stents or coronary-artery bypass surgery. N Engl J Med 2011;364:1016-26.
• Hammill, BG, Curtis, LH, Schulman, KA, et al. Relationship between cardiac rehabilitation and long-term risks of death and myocardial infarction among elderly medicare beneficiaries. Circulation 2010;121:63-70.
• Gao, WG, Lu, B, Sun, ML, et al. Comparison of atherosclerotic plaque by computed tomography angiography in patients with and without diabetes mellitus and with known or suspected coronary artery disease. Am J Cariol 2011;108:809-13.
• Chu, Z, Yang, Z, Dong, Z, et al. Characteristics of coronary artery disease in symptomatic type 2 diabetes patients: evaluation with CT angiography. Cardiovasc Diabetol 2010;9:74-81.
• Hulten, EA, Carbonaro, S, Petrillo, SP, et al. Prognostic value of cardiac computed tomography angiography. J Am Coll Cardiol 2011;57:1237-47.
• Taylor, AJ, Cerqueira, M, Hodgson, JM, et al. ACCF/SCCT/ACR/AHA/ASE/ASNC/NASCI/SCAI/SCMR 2010 appropriate use criteria for cardiac computed tomography. J Am Coll Cardiol 2010;57(22):1864-94.

You must be logged in to save to your library.

Jacc journals on acc.org.

• JACC: Advances
• JACC: Basic to Translational Science
• JACC: CardioOncology
• JACC: Cardiovascular Imaging
• JACC: Cardiovascular Interventions
• JACC: Case Reports
• JACC: Clinical Electrophysiology
• JACC: Heart Failure
• Current Members
• Campaign for the Future
• Become a Member
• Renew Your Membership
• Member Benefits and Resources
• Member Sections
• ACC Member Directory
• ACC Innovation Program
• Our Strategic Direction
• Our History
• Our Bylaws and Code of Ethics
• Leadership and Governance
• Annual Report
• Industry Relations
• Support the ACC
• Jobs at the ACC
• Press Releases
• Social Media
• Book Our Conference Center

Clinical Topics

• Chronic Angina
• Congenital Heart Disease and     Pediatric Cardiology
• Diabetes and Cardiometabolic     Disease
• Hypertriglyceridemia
• Invasive Cardiovascular Angiography    and Intervention
• Pulmonary Hypertension and Venous     Thromboembolism

Latest in Cardiology

Education and meetings.

• Online Learning Catalog
• Products and Resources
• Annual Scientific Session

Tools and Practice Support

• Quality Improvement for Institutions
• Accreditation Services
• Practice Solutions

Heart House

• 2400 N St. NW
• Washington , DC 20037
• Contact Member Care
• Phone: 1-202-375-6000
• Toll Free: 1-800-253-4636
• Fax: 1-202-375-6842
• Media Center
• Advertising & Sponsorship Policy
• Clinical Content Disclaimer
• Editorial Board
• Privacy Policy
• Registered User Agreement
• Terms of Service
• Cookie Policy

© 2024 American College of Cardiology Foundation. All rights reserved.

Log in using your username and password

• Search More Search for this keyword Advanced search
• Latest content
• Current issue
• Browse by collection
• BMJ Journals

You are here

• Volume 106, Issue 5
• Angina: contemporary diagnosis and management
• Article Text
• Article info
• Citation Tools
• Rapid Responses
• Article metrics

• http://orcid.org/0000-0003-4009-6652 Thomas Joseph Ford 1 , 2 , 3 ,
• http://orcid.org/0000-0002-4547-8636 Colin Berry 1
• 1 BHF Cardiovascular Research Centre , University of Glasgow College of Medical Veterinary and Life Sciences , Glasgow , UK
• 2 Department of Cardiology , Gosford Hospital , Gosford , New South Wales , Australia
• 3 Faculty of Health and Medicine , The University of Newcastle , Newcastle , NSW , Australia
• Correspondence to Dr Thomas Joseph Ford, BHF Cardiovascular Research Centre, University of Glasgow College of Medical Veterinary and Life Sciences, Glasgow G128QQ, UK; tom.ford{at}health.nsw.gov.au

https://doi.org/10.1136/heartjnl-2018-314661

Statistics from Altmetric.com

Request permissions.

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

• cardiac catheterization and angiography
• chronic coronary disease
• percutaneous coronary intervention
• coronary artery disease

Learning objectives

Around one half of angina patients have no obstructive coronary disease; many of these patients have microvascular and/or vasospastic angina.

Tests of coronary artery function empower clinicians to make a correct diagnosis (rule-in/rule-out), complementing coronary angiography.

Physician and patient education, lifestyle, medications and revascularisation are key aspects of management.

Introduction

Ischaemic heart disease (IHD) remains the leading global cause of death and lost life years in adults, notably in younger (<55 years) women. 1 Angina pectoris (derived from the Latin verb ‘angere’ to strangle) is chest discomfort of cardiac origin. It is a common clinical manifestation of IHD with an estimated prevalence of 3%–4% in UK adults. There are over 250 000 invasive coronary angiograms performed each year with over 20 000 new cases of angina. The healthcare resource utilisation is appreciable with over 110 000 inpatient episodes each year leading to substantial associated morbidity. 2 In 1809, Allen Burns (Lecturer in Anatomy, University of Glasgow) developed the thesis that myocardial ischaemia (supply:demand mismatch) could explain angina, this being first identified by William Heberden in 1768. Subsequent to Heberden’s report, coronary artery disease (CAD) was implicated in pathology and clinical case studies undertaken by John Hunter, John Fothergill, Edward Jenner and Caleb Hiller Parry. 3 Typically, angina involves a relative deficiency of myocardial oxygen supply (ie, ischaemia) and typically occurs after activity or physiological stress ( box 1 ).

Definition of angina (NICE guidelines) 32

Typical angina: (requires all three)

Constricting discomfort in the front of the chest or in the neck, shoulders, jaw or arms.

Precipitated by physical exertion.

Relieved by rest or sublingual glyceryl trinitrate within about 5 min

Presence of two of the features is defined as atypical angina.

Presence of one or none of the features is defined as non-anginal chest pain.

Stable angina may be excluded if pain is non-anginal provided clinical suspicion is not raised based on other aspects of the history and risk factors.

Do not define typical, atypical and non-anginal chest pain differently in men and women or different ethnic groups.

• Download figure
• Open in new tab
• Download powerpoint

Reappraisal of ischaemic heart disease pathophysiology. Distinct functional and structural mechanisms can affect coronary vascular function and frequently coexist leading to myocardial ischaemia. CAD, coronary artery disease.

We begin by classifying angina according to pathophysiology. We then consider the current guidelines and their strengths and limitations for assessing patients with recent onset of stable chest pain. We review non-invasive and invasive functional tests of the coronary circulation with linked management strategies. Finally, we point to future directions providing hope for improved patient outcomes and development of targeted disease-modifying therapy. The aim of this educational review is to provide a contemporary approach to diagnosis and management of angina taking into consideration epicardial coronary disease, microcirculatory dysfunction and coronary vasospasm.

Contemporary angina classification by pathophysiology

The clinical history is of paramount importance to initially establish whether the nature of the presenting symptoms is consistent with angina ( box 1 ). Indeed, recent data supports specialist physicians under-recognise angina in up to half of their patients. 10 Furthermore, contemporary clinical trials of revascularisation in stable IHD including the ISCHEMIA trial highlight the importance of good clinical history and listening to our patients to determine the nature and frequency of symptoms which helps to plan management. We propose a classification of angina that aligns with underlying aetiology and related management ( table 1 ).

• View inline

Classification of angina by pathophysiology

Angina with obstructive coronary artery disease

2018 ESC guidelines on myocardial revascularisation define obstructive CAD as coronary stenosis with documented ischaemia, a haemodynamically relevant lesion (ie, fractional flow reserve (FFR) ≤0.80 or non-hyperaemic pressure ratio (NHPR) (eg, iwFR≤0.89)) or >90% stenosis in a major coronary vessel ( table 1 ). There is renewed interest in NHPRs (iwFR, resting full-cycle ratio (RFR) and diastolic pressure ratio (dPR)) as data have emerged in support of numerical equivalency between these indices suggesting all can be used to guide treatment strategy. 11 Angina with underlying obstructive CAD allows symptom guided myocardial revascularisation (often with percutaneous coronary intervention (PCI)) and is effective in reducing ischaemic burden and symptoms (in many patients). Recent studies have served evidence that functional coronary disorders overlap and may contribute to angina even in patients with obstructive epicardial CAD. Dynamic changes in lesion or vessel ‘tone’ and propensity to vasoconstriction is important and may cause rest angina that is frequently overlooked in patients with obstructive CAD. 12 During invasive physiological assessment of ischaemia during exercise, Asrress et al showed that ischaemia developed at FFR averaging≈0.76 which is not often observed with adenosine induced hyperaemia. 13 This finding implies there are other important drivers of subendocardial ischaemia (myocardial supply:demand factors). Furthermore, it reinforces that angina is not synonymous with ischaemia or flow-limiting coronary disease (eg, abnormal FFR or NHPR). Coronary anatomy and physiology should not be considered in isolation but in the context of the patient.

Angina-myocardial ischaemia discordance

Although obstructive CAD or microvascular dysfunction may be present, the link between ischaemia and angina is not clearcut. The ‘ischaemic threshold’ (the heart rate-blood pressure product at the onset of angina) has intraindividual and interindividual variability. 14 Innate differences in vascular tone and endocrine changes (eg, menopause) may influence propensity to vasospasm while environmental factors including cold environmental temperature, exertion and mental stress are relevant. The large international CLARIFY registry highlighted the importance of symptoms, showing that angina with or without concomitant ischaemia, was more predictive of adverse cardiac events compared with silent ischaemia alone. 15 Other potential drivers of discordance between angina and ischaemia include variations in pain thresholds and cardiac innervation (eg, diabetic neuropathy).

Symptoms and/or signs of ischaemia but no obstructive coronary artery disease (INOCA)

Cardiologists are inclined to adopt a ‘stenosis centric’ approach to patient management; however, as clinicians we must appreciate that all factors are relevant, including coronary anatomy and function but systemic health and the psychosocial background ( figure 2 ). First, systemic factors including heart rate, blood pressure (and their product) and myocardial supply:demand ratio (Buckberg index) are relevant. 16 Reduced myocardial oxygen supply from problems such as anaemia or hypoxaemia should always be considered.

Contributing factors to myocardial ischaemia. The contributors to the physiological myocardial perfusion gradient and resultant ischaemia can be broken down at patient-level into systemic, cardiac and coronary factors. CAD, coronary artery disease; SEVR, subendocardial viability ratio.<Modified with permission from 47 >.

Second, coronary factors are well recognised but certain nuances are overlooked. In 2018, the first international consensus guidelines clarify that a definite diagnosis of MVA may be made in patients with angina with no underlying obstructive CAD, evidence of reversible ischaemia on functional testing and objective evidence of coronary microvascular dysfunction ( table 1 ). 17 ‘Probable MVA’ is defined by three of the above criteria. Coronary microvascular dysfunction may be structural (eg, small vessel rarefaction or increased media: lumen ratio) or functional (eg, endothelial impairment) and these disorders may coexist. Other coronary causes of INOCA include intramyocardial ‘tunnelled’ segments of epicardial arteries (myocardial ‘bridging’) who may have ischaemia on exercise. These segments are particularly susceptible to vasoconstriction due to endothelial impairment. 18 Coronary arteriovenous malformations are rare but may also cause of myocardial ischaemia. Vasospastic angina (‘Prinzmetal’s angina’) is typically described as recurrent rest angina with focal occlusive proximal epicardial often seen in young smokers with characteristic episodic ST segment elevation during attacks. Notably, the more common form of VSA is distal and diffuse subtotal epicardial vasospasm and is characterised by ST segment depression and may occur during exertion. Typical cardiac risk factors and endothelial impairment may be implicated. 19

The long-term (sometimes lifelong) burden of MVA and/or VSA on physical and mental well-being can be profound. Patients with these conditions commonly attend primary care and are repeatedly hospitalised with acute coronary syndromes, arrhythmias and heart failure driving up health resource utilisation, morbidity and reducing quality of life. 20 21

The third and final group of factors that drive ischaemia in patients with angina but without obstructive CAD include cardiac factors . These include left ventricular hypertrophy or restrictive cardiomyopathy where subendocardial ischaemia results impaired perfusion from arterioles penetrating deeper into myocardial tissue with shorter diastole, enhanced systolic myocardial vessel constriction and enhanced interstitial matrix. 22 Heart failure (with reduced or preserved ejection fraction) can lead to elevated left ventricular end diastolic pressure which reduces the diastolic myocardial perfusion gradient. Valvular heart disease (eg, aortic stenosis (AS)) is an important consideration in patients with INOCA. In AS, most experts support haemodynamic factors as the main cause of ischaemia, especially since symptoms and coronary flow reserve (CFR) improve immediately after valve replacement. 23 Patients with INOCA may have increased painful sensitivity to innocuous cardiac stimuli (eg, radiographic contrast) without inducible ischaemia. Furthermore, some affected patients have a lower pain threshold and tolerance to the algogenic effects of adenosine (thought to be the main effector of ischaemia mediated chest pain). 24

Gender differences and angina presentation

The WISE (Women’s Ischemia Syndrome Evaluation) study highlighted that over 2/3 of women with angina had no obstructive CAD and the majority of these had functional impairments in the coronary microcirculation associated with significant impairments in health-related quality of life. 25 Indeed, women have more non-obstructive CAD and functional IHD which are frequently overlooked and hence undertreated. 26 27 Over time and at different ages, women have a similar or slightly higher prevalence of angina than men across countries independent of diagnostic and treatment practices. 28 Different patterns of IHD may be anticipated to cause different angina symptoms between genders. Nonetheless, recent evidence moves the field away from the ‘male-typical, female-atypical’ model of angina towards a ‘gender continuum’ whereby the objective reports between men and women’s symptoms are more similar than treating physicians perceive. Interestingly, dyspnoea was a feature in around ¾ of angina presentations without any significant difference between the sexes. 29

Assessment: current guidelines

Assessment strategies in current major international guidelines focus on the detection of underlying obstructive CAD. European and American guidelines (ESC and ACC/AHA, respectively) favour a Bayesian approach whereby overall probability of obstructive CAD after testing is determined from pretest probability modified by the diagnostic test results. The ACC/AHA guidelines determine pretest risk from a modified Diamond Forrester model, 30 whereas the Europeans favour the CADC (Coronary Artery Disease Consortium) model which avoids overestimation seen with Diamond-Forrester and appears a more accurate assessment of pretest risk. 31 Both current guidelines stratify pretest risk into low, intermediate or high groups with use of non-invasive testing suggested in the intermediate group (ACC/AHA arbitrarily defined as 10%–90% or 15%–85% in ESC).

In stark contrast, the NICE CG95 2016 update ‘chest pain of recent onset: assessment and diagnosis’ discarded the Bayesian pretest risk assessment. NICE advocates first-line multidetector CT coronary angiography (CTCA) in all patients with typical or atypical chest pain ( box 1 ), those whose history does not suggest angina but who have ST changes or Q waves on a resting ECG. 32 Functional testing (eg, exercise stress echo or stress perfusion magnetic resonance—CMR) are relegated to second-line if CTCA is non-diagnostic or the clinical significance of known CAD needs clarified. Potential benefits of this approach include a much higher diagnostic accuracy for detection of atherosclerotic heart disease than functional testing which potentially carries the best long-term prognostic information for patients with CAD. 27 Extended 5-year outcomes from SCOTHEART showed a reduction in the combined endpoint of death from coronary heart disease or non-fatal myocardial infarction among the group randomised to CTCA compared with standard care (2.3% vs 3.9%; absolute risk reduction (ARR) 1.6% number needed to treat (NNT) 63). This effect was driven by better targeting of preventative therapies. The authors report that although overall prescriptions of preventive cardiovascular medications were only modestly increased (~10% higher) in the CTCA arm, changes in such therapies occurred in around one in four patients allowing more personalised treatment to patients with most coronary atheroma in the CT group.

These results should be considered in relation to design limitations of this trial. There was no control procedure (test vs no test), the threshold for prescribing preventive therapy with statins was 20%–30% likelihood of a CHD event in 10 years (much higher than many contemporary healthcare systems), CTCA was performed on top of treadmill exercise testing which has poor test accuracy in distinct patient groups, notably women, and the procedures were unblinded and open-label. Outcome reporting that is narrowly focused on CHD does not take account of other cardiovascular events, such as hospitalisation for arrhythmias and heart failure, which have implications for quality of life. In PROMISE, a ‘head-to-head’ trial of CTCA versus functional testing, there were no differences in health outcomes. 33 In the interests of providing patients and clinicians with a reliable and accurate test result, a strategy based on anatomical CTCA has fundamental limitations. SCOT-HEART identified that obstructive CAD affects the minority (one in four) patients presenting to the Chest Pain Clinic with known or suspected angina. This means that an anatomical test strategy with CTCA leaving the aetiology and treatment unexplained in the majority of affected patients, which becomes all the more relevant considering that anginal symptoms and quality of life are worse when CTCA is used. 34 Diagnostic options are enhanced by advances in technology and tests for the functional significance of CAD are now feasible, but at significant cost. 35 NICE guidelines state that HeartFlow FFR CT should be considered as an ‘option for patients with stable, recent onset chest pain who are offered CCTA as part of the NICE pathway on chest pain’. Using HeartFlow FFR CT may avoid the need for invasive coronary angiography and revascularisation; however, major randomised controlled trials are ongoing (eg, FORECAST study NCT03187639 ).

We support efforts to provide a definitive diagnosis for patients with ongoing angina symptoms after a ‘negative’ CTCA, initially using non-invasive ischaemia testing. Notably, the recent International Standardised Criteria for diagnosing ‘suspected’ MVA would be met in patients with symptoms of myocardial ischaemia, no obstructive CAD and objective evidence of myocardial ischaemia ( table 1 ). Invasive testing for diagnosis of MVA could be reserved for subjects with refractory symptoms and negative ischaemia testing or diagnostic uncertainty. The criteria for ‘definite MVA’ require the above AND objective evidence of microvascular dysfunction (eg, reduced CFR or raised microvascular resistance).

Limitations of current guidelines

There are limitations to the current NICE-95 guideline, not least the logistics and cost of service provision with an estimated 700% increase in cardiac CT required across the UK. 36 Importantly, what do we report to the majority of patients with anginal chest pain but no obstructive CAD on the CTCA? In fact, only 25% of patients had obstructive CAD and at 6 weeks based on the CTCA findings, 66% of patients were categorised as not having angina due to coronary heart disease. The possibility of false reassurance for the patients with angina and INOCA is an open question and may be one contributing factor for the lack of improvement in angina and quality of life in the CTCA group vs standard care. 34 We must strive to deliver patient-centred care, recognising that most patients seek explanation for their symptoms in combination with effective treatment options. 37 CTCA is an insensitive test for disorders of coronary vascular function, which may affect the majority of patients attending with anginal symptoms. Since the majority of affected patients have no obstructive CAD, and the majority of them are women, an anatomical strategy introduces a sex-bias into clinical practice, whereby a positive test result (obstructive CAD) is more likely to occur in men and a positive test for small vessel disease is less likely to occur in women. Furthermore, patient-reported outcomes including angina limitation, frequency and overall quality of life improve less after CTCA compared with standard care, notably in patients with no obstructive CAD. 34 Non-invasive functional testing with positron emission tomography (PET), echo and most recently stress perfusion CMR has diagnostic value for stratified medicine. Finally, stratification of patients using luminal stenosis severity on angiography overlooks the spectrum of risk associated with overall plaque burden and may miss functional consequences associated with diffuse but angiographically mild disease (particularly when subtending large myocardial mass).

Non-invasive functional testing includes myocardial perfusion scintigraphy, exercise treadmill testing (including stress echocardiography) or contrast-enhanced stress perfusion MRI depending on local availability. Novel pixel-wise absolute perfusion quantification of myocardial perfusion by CMR will likely improve the efficiency of absolute quantification of myocardial blood flow by CMR. 38 PET is the reference-standard non-invasive assessment of myocardial blood flow permitting quantitative flow derivation in mL/g/min. Clinically, PET-derived quantification of myocardial blood flow (MBF) can assist in the diagnosis of diffuse epicardial or microvascular disease; however, limitations include poor availability and exposure to ionising radiation. Non-invasive workup often provides important insights on coronary microvascular function and are reviewed in detail elsewhere. 39

With functional testing relegated to second-line testing, clinicians may forgo additional tests after a negative CTCA particularly in an era of fiscal restraint and if patients’ symptoms are viewed as atypical. One important group that will be disparately affected by an ‘anatomy first’ strategy are women—over half of all patients with suspected angina in the large prospective trials of CTCA are female. While the benefits of CTCA to diagnose CHD and prevent CHD events are similar in women and men, the large majority of patients undergoing CTCA do not have obstructive CAD potentially leading to misdiagnosis and suboptimal management in patients with INOCA. 33 Women, are most likely to have no obstructive CAD and their cardiac risk is associated with severely impaired CFR and not obstructive CAD. 40 Overall, there is growing awareness of sex-specific differences in coronary pathophysiology and potential for different patterns of CAD in women. This is a rapidly evolving fertile area for further research.

Invasive coronary angiography and physiological assessment

UK NICE guidelines suggest that invasive coronary angiography is a third-line investigation for angina when the results of non-invasive functional imaging are inconclusive. Patients with typical symptoms, particularly those in older age groups with higher probability of non-diagnostic CTCA scans, often proceed directly to invasive coronary angiography. During cardiac catheterisation, assuming that epicardial CAD is responsible for their symptoms, visual assessment for severe angiographic stenosis (>90%) is sufficient to establish significance and treatment plan for these patients. Two common pitfalls for visual interpretation of angiograms were recently highlighted by two coronary physiology pioneers Gould and Johnson. Using their quantitative myocardial perfusion database of over 5900 patients showing that occult coronary diffuse obstructive coronary disease or flush ostial stenosis may be both be overlooked on angiography and mislabelled as microvascular angina with suboptimal treatment. 41 The ischaemic potential of indeterminate coronary lesions (~40%–70% diameter stenosis) is best assessed using pressure-derived indices, such as FFR, and non-hyperaemic pressure ratios (NHPR: dPR, nstantaenous wave free ratio (iwFR) and others) to guide revascularisation decisions. However, as is the case with coronary angiography, these indices do not inform the clinician about disorders of coronary artery vasomotion.

Invasive tests of coronary artery function are the reference standard for the diagnosis of coronary microvascular dysfunction 17 and vasospastic angina ( table 1 ; figure 1 ). 42 Coronary microvascular resistance may be directly measured using guidewire-based physiological assessment during adenosine induced hyperaemia. Methods to assess this include using a pressure-temperature sensitive guidewire by thermodilution (index of microcirculatory resistance; IMR) or Doppler ‘ComboWire’ (hyperaemic microvascular resistance; HMR). These metrics have been the focus of a recent review article in Heart. 43 There are several other haemodynamic indices of microvascular function including instantaneous hyperaemic diastolic pressure velocity slope, wave intensity analysis and zero flow pressure. A detailed description of these parameters is out with the scope of this review. 41 Elevated coronary microvascular resistance (eg, IMR >25) carries prognostic utility in patients with reduced CFR but unobstructed arteries. Lee et al found over fivefold higher risk of adverse cardiac events in these subjects compared with controls with normal microvascular function. 44

CFR is the ratio of maximum hyperaemic blood flow to resting flow. CFR in the absence of obstructive CAD can signify impaired microvascular dilation. Lance Gould first introduced this concept almost 50 years ago but more recently proposed that CFR should be considered in the context of the patient and the hyperaemic flow rate. 41 The absolute threshold for abnormal CFR varies depending on the method of assessment, the patient population studies and the controversy reflects the dichotomous consideration of the continuous physiological spectrum of ischaemia. Abnormal CFR thresholds vary from ≤2.0 or ≤2.5 with more restrictive criteria for abnormal CFR (<1.6) being more specific for myocardial ischaemia and worse outcomes but at the cost of reduced sensitivity. On the other hand, studies of transthoracic Doppler derived CFR (which has less reproducibility) often use cut-offs of 2.5 with some observational evidence of worse outcomes in the INOCA population with CFR below this threshold. 45 The influence of rate-pressure product on resting flow and its correction for CFR determination should be considered.

Systolic endocardial viability ratio (SEVR) is a ratio of myocardial oxygen supply:demand derived from the aortic pressure-time integral (diastole:systole). However, it is well known that blood pressure, pulse and SEVR perturbations influence CFR more closely than microcirculatory resistance. Reduced CFR without raised microvascular resistance still portends increased cardiovascular risk 44 and may be a distinct subgroup with different drivers of ischaemia (eg, abnormal supply:demand systemic haemodynamic factors; figure 2 ). Alternatively, these patients may be at an earlier stage of disease prior to more established structural microvascular damage. Sezer et al showed the pattern of coronary microvascular dysfunction early in type II diabetes was driven by disturbed coronary regulation and high resting flow. 46 In longstanding diabetes however, elevated microvascular resistance was observed reflecting established structural microvascular disease. This process matches the paradox of microvascular disease in diabetic nephropathy where increased glomerular filtration rate (GFR) typifies the early stages of disease prior to later structural damage and reduction in GFR.

The third mechanism of microvascular dysfunction is inappropriate propensity to vasoconstriction of the small coronary arteries, typically this is assessed using intracoronary acetylcholine infusions as a pharmacological probe.

Rationale and benefit of invasive coronary function testing in INOCA

We contend that a complete diagnostic evaluation of the coronary circulation should assess structural and functional pathology. 47 The British Heart Foundation CorMicA trial provides evidence about the opportunity to provide a specific diagnosis to patients with angina using an interventional diagnostic procedure (IDP) when obstructive CAD is excluded by invasive coronary angiography. Consenting patients were randomised 1:1 to the intervention group (stratified medical therapy, IDP disclosed) or the control group (standard care, IDP sham procedure, results not disclosed). The diagnostic intervention included pressure guidewire-based assessment of FFR, CFR and IMR during adenosine induced hyperaemia (140 µg/kg/min). Vasoreactivity testing was performed by infusing incremental concentrations of acetylcholine (ACh) followed by a bolus vasospasm provocation (up to 100 µg). The diagnosis of a clinical endotype (microvascular angina, vasospastic angina, both, none) was linked to guideline-based management. The primary endpoint was the mean difference in angina severity at 6 months (as assessed by the Seattle Angina Questionnaire summary score—SAQSS) which was analysed using a regression model incorporating baseline score. A total of 391 patients were enrolled between 25/11/2016 and 11/12/2017. Coronary angiography revealed obstructive disease in 206 (53.7%). One hundred and fifty-one (39%) patients without angiographically obstructive CAD were randomised. The underlying abnormalities revealed by the IDP included: isolated microvascular angina in 78 (51.7%), isolated vasospastic angina in 25 (16.6%), mixed (both) in 31 (20.5%) and non-cardiac chest pain in 17 (11.3%). The intervention was associated with a mean improvement of 11.7 units in the SAQSS at 6 months (95% CI 5.0 to 18.4; p=0.001). In addition, the intervention led to improvements in the quality of life (EQ5D index 0.10 units; 0.01 to 0.18; p=0.024). After disclosure of the IDP result, over half of treating clinicians changed their diagnosis about the aetiology of their patients’ symptoms. There were no procedural serious adverse events and no differences in major adverse cardiac events (MACE) at 6 months. Interestingly, there were sustained quality of life benefits at one year for INOCA patients helped by correct diagnosis and linked treatment started at the index invasive procedure. 48 Future trials are anticipated to determine the wider external validity of this approach.

Medical therapy to prevent new vascular events should be considered and these include consideration of aspirin, ACE inhibitors (ACEi) and statins. The latter two agents have pleiotropic properties including beneficial effects on endothelial function and so may be helpful in treating coronary microvascular dysfunction. Sublingual glyceryl trinitrate tablets or spray should be used for the immediate relief of angina and before performing activities known to bring on angina.

Non-pharmacological

As with many cardiovascular diseases, lifestyle modification including risk factor control and patient education are key. Lifestyle recommendations are covered in detail in recent ESC guidelines. The adverse effect of angina on patient well-being and quality of life can be substantial. It is crucial that we assess for this and manage appropriately. After diagnosis with angina, cardiac rehabilitation can be useful to educate and build confidence. One useful patient led education aid is called the ‘Angina plan’. This tool is a workbook and relaxation plan delivered in primary care, which helps improve angina symptoms (frequency and limitation) while reducing anxiety and depression. 49 The ORBITA trial highlights the benefits of placebo effect and we support that the positive diagnosis may be therapeutic in itself. Angina symptoms are often subjective and multifactorial in origin, so patient education and validation of symptoms may facilitate further improvement.

Management: Non-obstructive CAD

Generic guidelines on angina management frequently overlooks the precision medicine goal whereby treatment is targeted to underlying pathophysiology. There is a lack of high-quality clinical trial data for treating microcirculatory dysfunction. The current article thus proposes a reasoned approach to management based on evaluation of pathophysiological mechanisms.

We contest that angina and INOCA are syndromes and not a precise diagnosis (akin to myocardial infarction with no obstructive CAD—MINOCA). As such, by stratifying treatment according to underlying pathophysiology, we may realise better outcomes for our patients.

Impaired coronary vasodilator capacity (reduced CFR)

Bairey Merz et al performed a randomised controlled trial of ranolazine in the WISE population. Notably, there was no net benefit effect on the INOCA population as a whole; however, in patients with reduced CFR (<2.5), there was a benefit suggestion of improved myocardial perfusion reserve index (MPRi) after established treatment. 50 Lanza and Crea highlight that subjects with reduced CFR might preferentially be treated with drugs that reduce myocardial oxygen consumption (eg, beta-blockers (BB)—for example, Nebivolol 1.25–10 mg daily). 51 There is accumulating evidence that long acting nitrates are ineffective or even detrimental in MVA. Lack of efficacy may relate to poor tolerability, steal syndromes through regions of adequately perfused myocardium and/or related to the reduced responsiveness of nitrates within the coronary microcirculation. 52 Furthermore, chronic therapy with nitrate may induce endothelial dysfunction and oxidative stress, predominantly via endothelin dependent pathways. 53

Increased microvascular constriction (structurally increased microvascular resistance or functional propensity to microvascular spasm)

Subjects with increased microvascular vasoconstriction may be treated with vasodilator therapies acting on the microcirculation. These include calcium channel blockers (CCB—for example, amlodipine 2.5–10 mg daily) or nicorandil (eg, 5–30 mg two times a day). Hyper-reactivity to constrictor stimuli resulting in propensity to microvascular spasm may be provoked by endothelial dysfunction. This was first described my Mohri et al over three decades ago with recent physiological studies suggesting treatment aimed at improving endothelial function (eg, ACEi, Ramipril 2.5–10 mg) may improve the microvascular tone and/or the susceptibility to inappropriate spasm. 54 55 A detailed discussion of all potentially therapeutic options for coronary microvascular dysfunction is beyond the scope of this article; however, a systematic review by Marinescu et al may be of interest to readers wishing further information. 56

Epicardial spasm (vasospastic angina)

The poor nitrate response or tolerance seen in MVA contrasts with patients with vasospastic angina, in whom nitrates are a cornerstone of therapy and BB are relatively contraindicated. 7 Dual pathologies (VSA with underlying microvascular disease) is increasingly recognised. A diagnosis of VSA facilitates treatment using non-dihydropiridine calcium antagonists (eg, diltiazem-controlled release up to 500 mg daily). Overall, CCB are effective in treating over 90% of patients. 57 High doses of calcium antagonists (non-dihydropiridine and dihydropyridine) may be required either alone or in combination. Unfortunately, ankle swelling, constipation and other side effects may render some patients intolerant. In these cases, long-term nitrates may be used with good efficacy in this group. In about 10% of cases, coronary artery spasm may be refractory to optimal vasodilator therapy. Japanese VSA registry data shows nitrates were not associated with MACE reduction in VSA, and importantly when added to Nicorandil were potentially associated with higher rates of adverse cardiac events. 58 Alpha blockers (eg, clonidine) may be helpful in selected patients with persistent vasospasm. In patients with poor nitrate tolerance the K+-channel opener nicorandil (5–10 mg two times a day) can be tried. Consider secondary causes in refractory VSA (eg, coronary vasculitis) and in selected patients with ACS presentations, coronary angioplasty may be considered as a bailout option.

Management: Obstructive CAD

Pharmacological.

Although NICE guidelines offer either BB or CCB first line, although we support BB initially because they are generally better tolerated ( table 2 ). 59 Long-term evidence of efficacy is limited between BB and CCB and there are no proven safety concerns favouring one or the other. Dihydropyridine calcium may be added to BB if blood pressure permits. NICE CG126 states third line options can be either added on (or substituted if BB/CCB not tolerated). These include nitrates (eg, isosorbide mononitrate 30–120 mg controlled release), ivabradine (eg, 2.5–7.5 mg two times a day), nicorandil (5–30 mg two times a day) or ranolazine (375–500 mg two times a day). These are all third line medications that can be used based and combined with BB and/or CCB depending on comorbidities, contraindications, patient preference and drug costs ( figure 3 ). The RIVER-PCI study found that anti-ischaemic pharmacotherapy with ranolazine did not improve the prognosis of patients with incomplete revascularisation after percutaneous coronary intervention. 60 This was a reminder that alleviation of ischaemia may not improve ‘hard’ endpoints in patients with chronic coronary syndromes but helps us to remain focused on improving their quality of life.

Angina pharmacotherapy

Empirical pharmacological treatments for patients with angina. ACEi, Angiotensin converting enzyme inhibitor; ASP, aspirin; BB, beta-blocker; Endo, endothelial; IVA, ivabradine; MVA, microvascular angina; NIC, nicorandil; NIT, nitrate; Obs CAD, obstructive coronary artery disease;, RAN, ranolazine; RF, risk factor.

Revascularisation

Recently revised 2018 ESC guidelines suggest that myocardial revascularisation is indicated to improve symptoms in haemodynamically significant coronary stenosis with insufficient response to optimised medical therapy. Patients’ wishes should be accounted for in relation to the intensity of antianginal therapy as PCI can offer patients with angina and obstructive CAD a reduced burden from polypharmacy. Angina persists or recurs in more than one in five patients following PCI and microvascular dysfunction may be relevant. Guidelines support consideration of revascularisation for prognosis in asymptomatic ischaemia in patients with large ischaemic burden (left main/proximal left anterior descending artery stenosis >50%) or two/three vessel disease in patients with presumed ischaemia cardiomyopathy (LVEF<35%).

Refractory angina is common in patients with complex CAD including those with previous coronary artery bypass grafting (CABG) and chronic total occlusions (CTOs). Over the last decade, vast strides in technique, training and tools have delivered major increases in the success of CTO PCI. These angina patients often have incomplete revascularisation with lesions or anatomy previously considered ‘unsuitable for intervention’ but now amenable to treatment by trained operators. A recent review article in Heart summarises non-pharmacological therapeutic approaches to patients with refractory angina including cognitive behavioural therapy (CBT), stellate ganglion nerve blockade, Transcutaneous Electrical Nerve Stimulation (TENS)/spinal cord stimulation and pain modulating antidepressants (eg, imipramine). 61 Of note, coronary sinus reducers deployed using a transcatheter venous system have shown early promise in clinical studies.

Future directions

Based on test accuracy, health and economic benefits, non-invasive and invasive functional tests should be considered a standard of care in patients with known or suspected angina, especially if obstructive CAD has been excluded by CT or invasive coronary angiography. Computational fluid dynamic modelling of the functional significance of CAD, notably with FFRct, is an emerging option and clinical trials, including FORECAST (ClinicalTrials.gov Identifier: NCT03187639 ) and PRECISE ( NCT03702244 ), are ongoing. The use of computational modelling as a diagnostic tool in patients with microvascular angina or coronary vasomotion disorders remains to be determined.

Systemic vascular abnormalities were recently highlighted in patients with INOCA potentially supporting a therapeutic role for targeted vascular therapy, for example, using selective endothelin-A receptor antagonists. 19 The MRC Framework for Stratified Medicine is applicable to patients with angina and we believe genetic testing with precision medicine holds future promise.

The optimal management of patients with known or suspected angina begins with establishing the correct diagnosis.Around one half of angina patients have no obstructive coronary disease; many of these patients have microvascular and/or vasospastic angina.Non-invasive assessment with CTCA is a sensitive anatomical test for plaque which assists in initial treatment and risk stratification. Anatomical imaging has fundamental limitations to rule in or rule out coronary vasomotion disorders in patients with symptoms and/or signs of ischaemia but no obstructive CAD (INOCA). Women are disproportionately represented in this group with MVA and/or VSA, the two most common causes of diagnoses. A personalised approach to invasive diagnostic testing permits a diagnosis to be made (or excluded) during the patients’ index presentation. This approach helps stratify medical therapy leading to improved patient health and quality of life. Physician appraisal of ischaemic heart disease (IHD) should consider all pathophysiology relevant to symptoms, prognosis and treatment to improve health outcomes for our patients. More research is warranted, particularly to develop disease modifying therapy.

ESC curriculum: stable CAD

Precipitants of angina.

Prognosis of chronic IHD.

Clinical assessment of known or suspected chronic IHD.

Indications for, and information derived from, diagnostic procedures including ECG, stress test in its different modalities (with or without imaging, exercise and stress drugs) and coronary angiography.

Management of chronic IHD, including lifestyle measures and pharmacological management.

Indications for coronary revascularisation including PCI/stenting and CABG.

Angina pectoris is a clinical syndrome occurring in patients with or without obstructive epicardial coronary artery disease.

Diagnostic testing in angina is symptom driven and so should provide patients and their physicians with an explanation for their symptoms and used to stratify management and offer prognostic insights.

Microvascular and/or vasospastic angina are common disorders of coronary artery function that may be overlooked by anatomical coronary testing, leading to false reassurance and adverse prognostic implications.

CME credits for Education in Heart

Education in Heart articles are accredited for CME by various providers. To answer the accompanying multiple choice questions (MCQs) and obtain your credits, click on the 'Take the Test' link on the online version of the article. The MCQs are hosted on BMJ Learning. All users must complete a one-time registration on BMJ Learning and subsequently log in on every visit using their username and password to access modules and their CME record. Accreditation is only valid for 2 years from the date of publication. Printable CME certificates are available to users that achieve the minimum pass mark.

Supplemental material

• Naghavi M ,
• Allen C , et al
• ↵ Cardiovascular Disease Statistics 2018 . London : British Heart Foundation , 2018 .
• Kligfield P
• ↵ Cine-coronary arteriography . Circulation . 227 East Washington Square, Philadelphia, PA 19106. : Lippincott Williams and Wilkins , 1959 .
• Kaski J-C ,
• Gersh BJ , et al
• Saraste A , et al
• Montalescot G ,
• Sechtem U ,
• Achenbach S , et al
• Corcoran D ,
• Peterson ED ,
• Dai D , et al
• Arnold SV ,
• Grodzinsky A ,
• Gosch KL , et al
• Park J , et al
• Asrress KN ,
• Williams R ,
• Lockie T , et al
• Garber CE ,
• Carleton RA ,
• Camaione DN , et al
• Greenlaw N ,
• Tendera M , et al
• Johnson NP ,
• Nitroglycerine JNP
• Camici PG ,
• Beltrame JF , et al
• Corban MT ,
• Prasad M , et al
• Rocchiccioli P ,
• Good R , et al
• Maddox TM ,
• Stanislawski MA ,
• Grunwald GK , et al
• Tavella R ,
• Tucker G , et al
• Raphael CE ,
• Parker KH , et al
• Pasceri V ,
• Buffon A , et al
• Kelsey SF ,
• Matthews K , et al
• Stanley B ,
• ↵ Ct coronary angiography in patients with suspected angina due to coronary heart disease (SCOT-HEART): an open-label, parallel-group, multicentre trial . The Lancet 2015 ; 385 : 2383 – 91 . doi:10.1016/S0140-6736(15)60291-4 OpenUrl
• Hemingway H ,
• Langenberg C ,
• Damant J , et al
• Kreatsoulas C ,
• Shannon HS ,
• Giacomini M , et al
• Diamond GA ,
• Forrester JS
• Bittencourt MS ,
• Polonsky TS , et al
• (NICE) NIfHaCE
• Douglas PS ,
• Hoffmann U ,
• Patel MR , et al
• Williams MC ,
• Shah A , et al
• Leipsic J ,
• Pencina MJ , et al
• Nicol EPS ,
• Roobottom C , British Society of cardiovascular Imaging/ British Society of cardiovascular computed tomography
• Richards T ,
• Coulter A ,
• Hsu L-Y , et al
• Szymonifka J ,
• Twisk JWR , et al
• Taqueti VR ,
• Cook NR , et al
• Beltrame JF ,
• Kaski JC , et al
• Aetesam-ur-Rahman M , et al
• Hwang D , et al
• Hachamovitch R ,
• Murthy VL , et al
• Kocaaga M ,
• Aslanger E , et al
• Berry C , B
• Sidik N , et al
• Robinson J , et al
• Bairey Merz CN ,
• Handberg EM ,
• Shufelt CL , et al
• Horowitz JD
• Kröller-Schön S , et al
• Koyanagi M ,
• Egashira K , et al
• Fearon WF ,
• Kobashigawa JA , et al
• Marinescu MA ,
• Löffler AI ,
• Ouellette M , et al
• Nishigaki K ,
• Yamanouchi Y , et al
• Takahashi J ,
• Takagi Y , et al
• Généreux P ,
• Iñiguez A , et al
• Sainsbury PA ,
• Jolicoeur EM

Supplementary materials

Supplementary data.

This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

• Data supplement 1

Twitter @tomjford

Contributors TJF devised and wrote the article and figures. CB edited and approved the final manuscript.

Funding British Heart Foundation (PG/17/2532884; RE/13/5/30177; RE/18/634217).

Competing interests CB is employed by the University of Glasgow which holds consultancy and research agreements with companies that have commercial interests in the diagnosis and treatment of angina (Abbott Vascular, AstraZeneca, Boehringer Ingelheim, GSK, Menarini, Opsens, Philips and Siemens Healthcare.)

Patient consent for publication Not required.

Provenance and peer review Commissioned; externally peer reviewed.

Read the full text or download the PDF:

• Get new issue alerts Get alerts

Secondary Logo

Journal logo.

Colleague's E-mail is Invalid

Your message has been successfully sent to your colleague.

Save my selection

A case report of right coronary artery agenesis diagnosed by computed tomography coronary angiography

Editor(s): NA.,

a IRCCS SDN, Naples

b Clinica Montevergine, Mercogliano, Italy.

∗Correspondence: Bruna Punzo, IRCCS SDN, via E. Gianturco 113, 80143, Naples, Italy (e-mail: [email protected] )

Abbreviations: AV = atrioventricular, CAD = coronary artery disease, CR = cinematic rendering, CTCA = computed tomography coronary angiography, LAD = left anterior descending artery, LCX = left circumflex artery, MIP = maximum intensity projection, MPR = multiplanar reformations, RCA = right coronary artery, SA = sinoatrial, VR = volume rendering.

How to cite this article: Forte E, Punzo B, Agrusta M, Salvatore M, Spidalieri G, Cavaliere C. A case report of right coronary artery agenesis diagnosed by computed tomography coronary angiography. Medicine . 2020;99:7(e19176).

The patient has provided informed consent for the publication of this case.

The authors have no conflicts of interest to disclose.

Funding source: Italian ministry of health; award id RICERCA CORRENTE 2018-2020

This is an open access article distributed under the Creative Commons Attribution License 4.0 (CCBY), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. http://creativecommons.org/licenses/by/4.0

Introduction: 

Single coronary artery is a rare condition characterized by the origin of a coronary that supplies the entire heart from a single coronary ostium.

Patient concerns: 

A 45-year-old woman with an altered exercise testing was addressed to a computed tomography coronary angiography (CTCA) to rule out coronary artery disease (CAD).

Diagnosis: 

CTCA examination showed the absence of the right coronary artery (RCA). The left anterior descending artery and the left circumflex artery (LCX) presented regular origin and course and LCX provided the posterior interventricular artery and the posterolateral artery.

Interventions: 

As CTCA highlighted the absence of potentially life-threatening features related to coronary anomaly, no surgical treatment was advised.

Outcomes: 

The patient was dismissed, kept under pharmacological control and monitored over time.

Conclusion: 

CTCA is the first-choice imaging modality in patients with ECG abnormalities properly allowing the differential diagnosis between CAD and congenital heart disease.

1 Introduction

Congenital agenesis of the right coronary artery (RCA) is a coronary anomaly characterized by a single coronary ostium from which the whole coronary tree takes origin. Coronary anomalies are often incidental findings in patients undergoing coronary angiography for coronary artery disease (CAD) exclusion. Cardiovascular computed tomography, thanks to its high spatial and contrast resolution, constitutes an excellent tool for the differential diagnosis between CAD and congenital anomalies providing typical findings. [1]

In this study, we report the case of a patient with congenital absence of RCA who performed a computed tomography coronary angiography (CTCA) for CAD suspicion.

2 Case report

An asymptomatic 45-year-old woman was referred to our institution with an exercise testing suggestive for myocardial ischemia (ST segment depression), according to AHA guidelines. [2] She was not under pharmacological treatment. Echocardiography and biochemical parameters were within the normal limits. Afterwards, a CTCA was planned to rule out CAD. The examination was carried out by a 128-slice computed tomography (Revolution GSI, GE Healthcare) with 0.35 second rotation time, prospective ECG triggering and 40 mm total collimation width. The angiographic scan was preceded by a baseline acquisition for calcium score evaluation. Therefore, 100 mL of high concentration contrast medium (Iomeprol 400 mg I/mL, Iomeron 400, Bracco, Milan, Italy) were injected at 5 mL/s and bolus tracking technique was used to synchronize the start of the acquisition with contrast bolus arrival. Images were reconstructed using the GE advantage workstation 4.6 (GE Healthcare) and Cinematic Rendering Prototype (version 1.0.1, Syngo.VIA Frontier Platform, Siemens Healthineers).

The phase with least residual motion was utilized for further evaluation. Maximum intensity projection (MIP), multiplanar reformations (MPR), 3D volume rendering (VR) and cinematic rendering (CR) were generated ( Figs. 1 and 2 ).

The examination showed the absence of any artery arising from the right sinus of Valsalva. The left anterior descending artery (LAD) and the left circumflex artery (LCX) presented regular origin and course with no evidence of CAD (Calcium score = 0). LCX provided a wide obtuse marginal branch, a filiform posterior interventricular artery and a noticeable posterolateral artery running in the inferior atrioventricular groove reaching the crux cordis up to the right atrioventricular sulcus. The LAD provided an acute margin branch and the conus artery while the sinoatrial nodal artery originated from proximal LCX. In light of imaging findings, the surgical intervention was not kept under consideration. Therefore, the patient was kept under pharmacological control by anti-platelets and monitored over time without any surgical intervention. Moreover, she was advised to avoid heavy exercise and to observe a proper nutrition and a healthy lifestyle to prevent the onset of coronary atherosclerosis.

3 Discussion

Single coronary artery belongs to the broad spectrum of coronary artery anomalies and constitutes a rare condition where the entire coronary arterial system arises from a solitary ostium. [3] For single coronary artery anomalies, the Lipton classification, revised by Yamanaka and colleagues, [4] is commonly used and includes three groups according to the site of origin and the vessel course. [5] Single coronary artery has a low prevalence in general population ranging from 0.014% to 0.066%. [5,6]

Coronary congenital absence is thought to be caused by a defect of coronary development during the embryonic period resulting in a coronary artery anomaly [1] and it may be also associated with other congenital heart disease. [3] This condition has to be differentiated by congenital ostial atresia, which represents a separate entity characterized by partial or total arterial orifice occlusion and corresponding proximal coronary hypoplasia. [7,8]

Patients may be asymptomatic or present with myocardial ischemia, acute coronary syndrome, syncope, ventricular fibrillation, or sudden death. [6] Many authors [9,10] consider some mechanisms explaining myocardial ischemia in single coronary artery, such as the coronary steal phenomenon due to the abnormal vessel or microvascular damage, and slow controlled ischemia caused by long travel distance of abnormal coronary artery. Shirani et al [11] reported that 15% of patients with a single coronary artery had myocardial ischemia due to the direct consequence of the coronary anomaly.

As for ECG alterations subsequent to single coronary artery, Yan et al [6] reported that the ECG manifestations of a patient with single left coronary artery may vary from no changes to various findings such as nonspecific ST-T wave changes to supraventricular arrhythmia. To a certain extent this may be explained by the blood supply to sinoatrial (SA) and atrioventricular (AV) node. While RCA normally supply these nodes, it is done so by a single left coronary artery and its branches. Lack of adequate blood supply from this anomaly may lead to ischemia in SA and AV node with eventual fibrosis and dysfunction which might be manifested in ECG varying from normal to abnormal readings related to ischemia or arrhythmia.

The relationship between the clinical manifestations and congenital absence of the RCA is still unknown. It is presumed that during an early age, absence of risk factors such as atherosclerosis, hypertension and diabetes may be the reason behind the lack of clinical symptoms. Patients are symptomatic only when the disease of coronary artery begins to occur.

Multidetector computed tomography has gained a crucial role in cardiovascular disease diagnosis and the recent technological improvements have widened its application allowing more patients to be acquired with contextual radiation dose reduction. [12] The updated 2016 NICE guideline indicates CTCA as the first-line investigation in all patients with atypical or typical angina symptoms or those who are asymptomatic with suggested ECG changes for ischemia. [13] At the same time, CTCA is increasingly applied in congenital heart disease diagnosis and follow-up providing an excellent visualization of systemic venous, abnormalities, extracardiac blood vessels, and coronary arteries. [14] Invasive coronary angiography is considered the gold standard for the diagnosis and treatment of coronary diseases even if it shows the vessel lumen not providing information about the vessel wall or the surrounding structures. [1] Moreover, when a coronary absence occurs, the operator is faced with a more laborious procedure when trying to catheterize a non-present coronary ostium. [6]

In our case, the results of CTCA clearly demonstrate that myocardial ischemia is not provoked by coronary artery disease. Moreover, CTCA revealed RCA absence and displayed a markedly dominant LCX giving off branches to the right atrium and the right ventricle (Group I Lipton anomalies, L-I pattern). Even though this can be considered a potentially serious coronary anomaly, [4] CTCA underlined the absence of angina related features such as acute angle take-off, slit-like ostium, ostial tissue flaps, coronary intussusception, or aberrant course between the aorta and the pulmonary artery.

As for patient management, currently, there is no standardized procedure or guideline from the evidence-based medicine for the treatment of congenital absence of the RCA. Choice of treatment may include either conservative treatment with anti-platelets, lipid-lowering, anti-hypertensive therapy, etc or with interventional therapy such as coronary artery revascularization, pacemaker implantation and other cardiac surgical procedures. [6]

Angelini [15] investigated coronary artery anomalies and proposed a diagnostic protocol for adult patients who are at risk for coronary artery anomalies where percutaneous transluminal coronary angioplasty or surgery are considered only in case of severe narrowing due to intramural course. Asymptomatic right anomalous origin of the coronary artery arising from the opposite sinus is even more challenging, given its weaker association with sudden cardiac death. In a significant number of asymptomatic cases, intervention may not be warranted, particularly if there is no evidence of ischemia in the patient. [16]

In conclusion, CTCA confirmed to be the choice technique for coronary visualization and CAD exclusion, reducing the number of inconclusive invasive procedures.

Author contributions

Conceptualization: Ernesto Forte.

Resources: Ernesto Forte.

Supervision: Carlo Cavaliere.

Visualization: Bruna Punzo, Marco Agrusta, Marco Salvatore, Gianluca Spidalieri.

Writing – original draft: Ernesto Forte.

Writing – review & editing: Ernesto Forte, Bruna Punzo.

• Cited Here |
• Google Scholar

coronary agenesis; coronary anomalies; coronary computed tomography angiography; right coronary artery

• + Favorites
• View in Gallery

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Open access
• Published: 12 June 2024

Enhancing the diagnosis of functionally relevant coronary artery disease with machine learning

• Christian Bock   ORCID: orcid.org/0000-0002-0701-5868 1 , 2   na1 ,
• Joan Elias Walter 3 , 4 , 5   na1 ,
• Bastian Rieck   ORCID: orcid.org/0000-0003-4335-0302 1 , 2 , 6   na1 ,
• Ivo Strebel 3 , 4 ,
• Klara Rumora 3 , 4 ,
• Ibrahim Schaefer 3 , 4 ,
• Michael J. Zellweger 3 , 4 ,
• Karsten Borgwardt   ORCID: orcid.org/0000-0001-7221-2393 1 , 2 , 7   na2 &
• Christian Müller   ORCID: orcid.org/0000-0002-1120-6405 3 , 4   na2  

Nature Communications volume  15 , Article number:  5034 ( 2024 ) Cite this article

514 Accesses

6 Altmetric

Metrics details

• Machine learning
• Translational research

Functionally relevant coronary artery disease (fCAD) can result in premature death or nonfatal acute myocardial infarction. Its early detection is a fundamentally important task in medicine. Classical detection approaches suffer from limited diagnostic accuracy or expose patients to possibly harmful radiation. Here we show how machine learning (ML) can outperform cardiologists in predicting the presence of stress-induced fCAD in terms of area under the receiver operating characteristic (AUROC: 0.71 vs. 0.64, p  = 4.0E-13). We present two ML approaches, the first using eight static clinical variables, whereas the second leverages electrocardiogram signals from exercise stress testing. At a target post-test probability for fCAD of <15%, ML facilitates a potential reduction of imaging procedures by 15–17% compared to the cardiologist’s judgement. Predictive performance is validated on an internal temporal data split as well as externally. We also show that combining clinical judgement with conventional ML and deep learning using logistic regression results in a mean AUROC of 0.74.

Similar content being viewed by others

Predicting long-term time to cardiovascular incidents using myocardial perfusion imaging and deep convolutional neural networks

Time and event-specific deep learning for personalized risk assessment after cardiac perfusion imaging

Machine learning and deep learning to predict mortality in patients with spontaneous coronary artery dissection

Introduction.

Coronary artery disease (CAD) is the leading cause of death worldwide 1 , 2 , 3 . High mortality and morbidity rates, paired with the availability of highly effective and cost-efficient prevention and treatment measures, underline the importance of early risk stratification of patients with suspected CAD. CAD may be clinically silent for decades or become functionally relevant (fCAD) by causing symptoms of myocardial ischaemia that impact the quality of life and potentially result in significant or adverse cardiac events such as premature death or nonfatal acute myocardial infarction (AMI) in the further course. Therefore, detection strategies should focus on fCAD to maximise patient benefit. Unfortunately, rapid, easy, and safe rule-out of fCAD remains a major unmet clinical need. The practical utility of current screening techniques is limited by either unfavourable diagnostic accuracy, as in the case of exercise electrocardiography stress testing, or by their obtrusive nature and high costs, as in the case of functional non-invasive imaging such as myocardial perfusion imaging (MPI) or anatomical non-invasive evaluation such as coronary computed tomography angiography 4 , 5 , 6 , 7 , 8 , 9 . While these dedicated cardiac imaging techniques can benefit many patients, they appear to be increasingly employed improperly in patients with a low pre-test probability of fCAD 5 , 6 , 7 . Considering the large population at risk, as well as the available prevention and treatment options, a clinical tool that enables effective, efficient, and safe detection of fCAD can improve patient outcomes while reducing the burden on patients as well as health care costs.

The automated detection of cardiac events has a long history 10 , 11 , and traditionally employed methods rely on quantifying ECG changes such as ST-segment elevation/depression, T-wave abnormalities, or other morphological anomalies of the QRS complex. However, a significant drawback of these methods is their reliance on ECG delineation algorithms that locate the segments a heartbeat is composed of. Delineation results can be inaccurate 10 for abnormal heartbeats, thus substantially limiting their utility for at-risk patients. Over the last few years, deep learning (DL) emerged as a powerful tool to build classification systems from ECG signals that do not require engineering QRS complex features 12 , 13 . Particularly in detecting different cardiac arrhythmias, the classification performance of DL systems reached the point of cardiologist-level accuracy 14 , 15 . While the potential of DL has been investigated in the context of cardiac stress testing 16 , 17 , 18 , 19 , 20 , previous work has the following drawbacks: (1) usage of a large number of variables which exacerbates model transferability, (2) reliance on summary statistics computed from automated ECG delineation or automated and less accurate outcome definitions, (3) lack of comprehensive performance evaluations on diverse subcohorts, and (4) lack of external validation. Lastly, ours is the first study investigating the benefit of collaborative machine learning in predicting abnormal myocardial perfusion.

Recent cardiology clinical practice guidelines 8 , 21 discouraged the sole use of stress ECG testing due to low diagnostic accuracy and unacceptable false negative and positive rates. However, given its wide availability, ease of use, and low cost, stress ECG testing remains commonly performed, which demands methods to use data acquired during stress testing more effectively. In addition, a stress ECG contains a plethora of information that cannot be included in routine clinical assessment (such as subtle morphological changes over time) but can serve as clinically relevant input for a DL system. At the same time, conventional machine learning based on static clinical variables alone has been shown to be at least as powerful as more complex deep neural networks in the healthcare setting 22 , 23 , 24 .

Thus, our aim is to derive and validate two different machine learning models in a heterogeneous patient population with a wide range of pre-test probabilities, namely (1) an ensemble learning model based on basic available clinical information, and (2) a deep learning model based on the aforementioned non-sequential variables as well as the ECG signals obtained during stress testing. We compare these models to the clinical assessment of cardiologists after stress testing. Furthermore, to extend their possible scope, the models were also trained and evaluated in patients who are usually excluded from stress ECGs and compared with the cardiologist’s clinical assessment after pharmacological testing.

Data collection, label generation, and robustness

Panel a of Fig.  1 illustrates our data generation workflow. We collected stress test ECG data from 3522 consecutive adult patients who underwent a standard 25 rest/stress myocardial perfusion single-photon emission computed tomography (SPECT) protocol at a tertiary hospital as part of the BASEL VIII study (NCT01838148). Patients were referred with symptoms possibly related to inducible myocardial ischaemia and clinical suspicion of stable coronary heart disease. If a patient was not able to reach their target heart rate, a pharmacological protocol with either adenosine or dobutamine was initiated by the treating clinician. Individuals for whom stress test by bicycle ergometry was not possible were put on a pharmacological protocol from the start. To compare the algorithmic approaches with expert judgement, the treating cardiologist performed a clinical assessment before and after stress testing: considering all available medical information such as (cardiac) history, relevant symptoms, risk factors, (stress) ECG, prior imaging and more, they indicated the probability of the presence of fCAD on a visual analogue scale (VAS) from 0% to 100% 26 , 27 , 28 , 29 . Representing clinical practice, adjudication of functionally relevant CAD was not formally blinded for stress ECG results or demographics and was performed centrally by an expert team composed of a nuclear medicine physician and a cardiologist assessing myocardial perfusion scans. Furthermore, whenever available, adjudication was refined with coronary angiography and fractional flow reserve assessment. Of the 3522 eligible patients who provided written informed consent, 701 (20%) patients underwent coronary angiography within 3 months, with 30 (0.9%) patients being reclassified to the fCAD group and 74 (2.1%) being reclassified to the non-ischaemic group. The VAS score the treating cardiologist provides after the stress test but before they get access to the imaging results represents the cardiologist baseline in our study. In practice, this can be interpreted as an indicator as to whether the cardiologist would recommend a follow-up examination with advanced imaging.

a Data acquisition: We highlight the three primary subgroups of exercise stress testing: ① patients who complete the bicycle exercise stress test, ② patients not able to exercise on the bicycle, and for whom a pharmaceutical protocol is used at the beginning of the stress test, and ③ patients starting on the bicycle but need pharmacological support to reach their target heart rate. Doctors perform myocardial perfusion scans at rest (rest MPS), and at the target heart rate (stress MPS). Myocardial perfusion is quantified by the myocardial perfusion scan summed rest score (MPSSR score), and the MPS summed stress score (MPSSS score). The cardiologist estimates the probability of a functionally relevant CAD (fCAD) before and after the stress test (Pre/Post-Test CAD Probability in the figure). The binary label indicating the presence of fCAD (yellow box) is adjudicated by considering the stress test results and additional relevant clinical parameters. b Data Preprocessing: Following smoothing and outlier removal, time series that serve as input to the neural network are constructed by joining short subsequences from different phases of the stress test. For this, 2 s from the pre-stress phase, 6 s from the stress phase, and 2 s from the recovery phase are sampled and concatenated multiple times for a single patient (green, orange, and purple sequences). x -axes represent time in seconds. c Machine Learning: For our neural network approach (CARPE ECG ), these 2-6-2 sequences are fed into a residual neural network (ResNet). In parallel, the patient’s static clinical data are processed by a 2-layer feedforward network. Four subnetworks are trained on three auxiliary tasks (i.e., MPSSR & MPSSS score as well as stress type prediction) and one main task (fCAD prediction). We average predictions of the main task over all 2-6-2 sequences per patient. Purple arrows in front of each task indicate the direction of the learning signal. The same clinical variables as for CARPE ECG are used to train a random forest classifier (CARPE Clin. ); nodes are coloured to enhance legibility. We combine both predictions with the cardiologist’s judgement in a logistic regression model (CARPE Coll. ).

The data set was split into a development (75%) and a held-out test set (25%). All patients in the development set enrolled in the study from Jan. 2010 through Dec. 2014; the held-out test set contains patients who enrolled from Dec. 2014 through May 2016. It was only released and used once the models’ parameters were fixed. Thus, high predictive performance on the held-out test set indicates the robustness of our system’s generalisation capability with respect to a temporal shift of the data 30 , paving the path towards subsequent real-world applications. Lastly, we use external data from two Israeli medical centres to validate our system on 916 consecutive patients referred for SPECT MPI testing, whose ECG signals were obtained by treadmill stress test. This evaluation scenario is designed to exemplify the ability of both computational approaches to generalise to patients from unseen institutions, new modalities, and highlight their behaviour under distributional shifts. Given an fCAD prevalence of 7.5% in the external data set, our approach based on clinical data alone (AUROC: 0.75 ± 0.004, AUPRC: 0.19 ± 0.01) is outperformed by our deep neural net using ECG time series and clinical variables (AUROC: 0.80 ± 0.01, AUPRC: 0.28 ± 0.01). Please refer to the Method section for more details on data splitting and distributional shifts in the external validation data.

Development of an ensemble predictor and a multi-task neural network for functionally relevant CAD prediction

The ability to learn from raw sequential data (i.e., time series) makes neural networks a popular approach for healthcare applications. However, conventional machine learning (ML) has shown to be at least as powerful as deep learning in the clinical context 22 , 23 , thus creating opportunities for low-cost deployments that do not require specialised hardware. Therefore, we will also compare their performance to deep learning models. To this end, we select a small set of eight non-sequential, easy-to-access variables on which we train four conventional ML methods (i.e., decision trees, random forests 31 , logistic regression, support vector machines 32 ). These variables include age, weight, biological sex, height, heart rate at rest, systolic and diastolic blood pressure, and presence of a previous CAD. The best-performing approach (a random forest) was selected via 5-fold cross-validation. We refer to all developed predictors as Coronary ARtery disease PrEdictor (CARPE). Based exclusively on clinical data, we refer to the random forest model as CARPE Clin. . Additionally, we develop a neural network approach, CARPE ECG , that uses the aforementioned non-sequential variables and the ECG signal, as illustrated in panel c of Fig.  1 . We trained CARPE ECG via a multi-task learning 33 (MTL) architecture with residual layers 15 , 34 at its core using the torchmtl 35 package. MTL uses so-called auxiliary tasks (i.e., prediction targets) related to a main task (e.g., fCAD prediction). These domain-specific inductive biases ensure improved and robust predictive performance on the main task 36 . As shown in Fig.  1 , we train CARPE ECG on three auxiliary tasks (blue boxes), two of which (MPSSS and MPSRS) quantify the heart’s perfusion capabilities without and under stress, respectively. The third auxiliary task is to predict whether a patient received any pharmacological support to perform the stress test. Each auxiliary task impacts performance on the main task differently. Their respective importance weights were selected in a grid search on the three best-performing leads (see Supplementary Fig.  5 and Supplementary Table  5 ). To gain insights into the importance of static features and ECG segments, we used SHAP (SHapley 37 Additive exPlanation) values 38 .

Finally, we combine predictions from the ensemble model and deep learning approach with the cardiologist’s post-test judgement by training a new logistic regression model on all three scores from the training set. This way, we leverage the experience and domain knowledge of the cardiologist while adding the potential to benefit from supervised learning techniques. We believe that, in practice, such a collaborative approach has the highest chances of being accepted in a clinical setting not only because it reaches the highest diagnostic performance (see Supplementary Table  6 ) but also because the cardiologist is an integral part of the score generation (in fact they are required to provide a VAS score after stress testing). Nevertheless, the use of logistic regression to enhance diagnostic accuracy does not ensure or directly translate to clinical utility. The precise impact on patient risk stratification needs to be assessed separately.

Machine learning can be used to reduce unnecessary perfusion imaging

The prevalence of centrally adjudicated fCAD was 32.9% in the full study cohort and 28% in the held-out test split. Figure  2 depicts the diagnostic performance of our machine learning approaches, the cardiologist’s assessment after stress testing, and a computational approach that uses the ECG’s ST-segment depression 39 , 40 as an indication of the presence of fCAD (see Methods for a detailed description) on the held-out data set. We show receiver operating characteristic (ROC) and precision-recall curves in the first row. Standard deviations shown as envelopes were obtained using bootstrapping, as detailed in the Methods section. Regarding the mean area under the ROC curve, we observe that CARPE ECG (0.71) and CARPE Clin. (0.70) outperform both the ST-depression algorithm (0.58) and the cardiologist (0.64). In regions of high specificity, CARPE Clin. drops below the sensitivity of the cardiologist, while CARPE ECG reaches comparable predictive performance (see inline plot). At the other extreme of the ROC curve, i.e., at high sensitivity values, both machine learning approaches consistently lead to a higher specificity than the cardiologist’s judgement (see inline plot).

ROC and PR-curve. Predictive performance of our deep learning-based approach (CARPE ECG ), a random forest based on clinical data (CARPE Clin. ), the cardiologist, and ST depression in terms of mean performance ± standard deviation (envelopes) over n  = 25 bootstrap draws. The upper plots show that both machine learning approaches outperform the cardiologist in terms of area under the receiver operating characteristic and precision-recall curve. In regions of high specificity (inline plot), the neural network is on par with the cardiologist while CARPE Clin. exhibits worse performance. Both machine learning methods outperform the cardiologist’s judgement in regions of high sensitivity (inline plot). Decision Curve : First row: Net benefit 43 plot for CARPE ECG (green), CARPE Clin. (orange), the cardiologist (purple), a myocardial perfusion scan (MPS) for no patient (black), and MPS for all patients (dashed grey). CARPE Coll. is not shown as it is visually indistinguishable from CARPE ECG . Net benefit puts both benefit and harm on the same scale. In our case, we consider harm to be inflicted by performing an unnecessary MPS. At a decision threshold of 5%, all approaches lead to a similar net benefit. At the second threshold of 15%, CARPE Clin. and the cardiologist demonstrate a net benefit similar to performing MPS on all patients, with CARPE ECG leading to a higher net benefit. Second row: Potential MPSs avoided compared to the cardiologist’s strategy: While the conventional ML model and deep learning avoid the approximately same number of MPSs at the decision threshold of 5% (11.5% and 12.8%, respectively), the gap increases at the pre-MPS threshold of 15% (15.3% and 5.3%, respectively). Envelopes in both rows show 95% confidence intervals around the mean over n  = 25 bootstrap draws. Source data are provided as a Source Data file.

Decision curves 41 (rows two and three in Fig.  2 ) overcome the drawbacks of conventional performance evaluations and calibration analyses 42 by focusing on a predictor’s clinical value. The concept of net benefit quantifies the trade-off between diagnosing sick patients and preventing healthy patients from being exposed to harmful testing procedures 43 . For a specific decision threshold probability of a diagnostic tool, a larger net benefit indicates a greater number of true positive predictions without an increase in the rate of false positives and, conversely, a greater number of true negative predictions without an increase in false negatives. Figure  2 shows a decision curve analysis in the second and third row with pre-test rule-out cutoffs (dotted red) as advocated for in European and US-American guidelines 8 , 21 , which consider probability thresholds between 5-15% for further non-invasive imaging. The European guideline, for instance, considers non-invasive testing in patients with a probability >15% as most beneficial and testing in patients with 5–15% as potentially beneficial. Our machine learning models lead to a higher net benefit than the cardiologist’s assessment at all thresholds. Notably, at the threshold of 15%, relying on the cardiologist’s judgement is worse (in terms of net benefit) than performing myocardial perfusion imaging on all patients demonstrating the value of an ML-based method.

Table  1 offers a detailed decision curve analysis, showing sensitivity, negative predictive value (NPV), and percentage of avoided myocardial perfusion imaging compared to the cardiologist’s judgement at three probability cut-off values. We also show the percentage of patients who received a score below the cutoff threshold to enable a meaningful interpretation of sensitivity values. The highest fraction of MPIs, i.e., almost 25 per 100 patients, could be avoided at a decision threshold of 10% by using CARPE Coll. as a risk stratification method due to risk-overestimation by the cardiologist. That being said, cutoff thresholds should not be chosen to optimise diagnostic performance, but they represent the cardiologist’s minimum probability of disease at which an intervention would be warranted 42 . In other words, if a cardiologist holds the belief that missing a patient who suffers from fCAD is nine times worse than performing an unnecessary MPI, a model’s performance should be assessed at the 10% cutoff.

Evaluating CARPE ECG as a predictive model on all patients of the held-out test set (at the 15% decision threshold) shows the potential to reduce perfusion imaging by 15.3% (see Table  1 ) without increasing the rate of false negatives. This number increases to 17.3% when using CARPE Coll. . We observe similar behaviour in patients without a CAD history. At the 5% threshold (i.e., if a physician considers it 19 times worse to miss an fCAD diagnosis than to perform an unwarranted MPI), ML can be used to avoid 10.8% of the imaging ordered by a cardiologist. For patients with CAD history, the decision thresholds of 5% and 10% lead to a particularly small number (<1% or none) of patients for which fCAD can be ruled out, which inflates sensitivity and NPV of CARPE ECG and CARPE Coll. . This inflation is particularly pronounced in CARPE Clin. (see Supplementary Fig.  6 ) which is therefore not shown here. Overall, these results demonstrate the potential clinical utility of the proposed methods to reduce potentially unwarranted MPIs.

Subcohort analysis: machine learning performs particularly well on younger patients

Trustworthiness and interpretability are of fundamental importance in the development of risk stratification models in cardiology 44 . Identifying (sub)cohorts of the population for which the model performs particularly well or poorly is crucial. To address the issue of trust, we evaluate our models’ performances on a variety of subcohorts that are important in the context of exercise stress testing. Regarding interpretability, we perform an analysis of SHAP values 38 on the population level, and a case study to better understand the impact feature values and ECG segments have on the predicted scores.

Clinically significant subgroups include patients who underwent exercise stress testing versus patients who required pharmacological testing as well as patients without a prior history of CAD versus patients with a known history of CAD; the odds of suffering from fCAD are significantly increased ( p  = 2.26E-40, two-sided Fisher’s exact test, test statistic = 2.64) for patients with previous CAD (OR: 2.64, 95% CI: 2.28–3.05) over the whole cohort. To obtain a more detailed performance breakdown, we also stratify the data by sex and age. Diagnostic performances of all approaches and subcohorts of the held-out test set are shown in Fig.  3 and Supplementary Table  6 . For comparison, the performance of the CAD consortium model 45 and the currently used ESC pre-test probabilities for obstructive coronary artery disease 8 , 9 , both based on age, sex, and the nature of symptoms, is shown in patients without known coronary artery disease. First, we assess the performance of individual machine learning methods before discussing their combination with the cardiologist’s judgement. Deep learning outperforms the cardiologist in terms of both AUROC (significant performance increase in 6/10 subcohorts) and AUPRC (significant performance increase in 4/10 subcohorts), while CARPE Clin. exceeds the human baseline in 5/10 strata in terms of AUROC and 2/10 subcohorts in terms of AUPRC. The central plot in panel a of Fig.  4 helps explain this performance discrepancy: The conventional ML model relies more than the neural network on the CAD history and sex variable as visually observable by the large gap between the highest negative and the lowest positive SHAP value for each variable. Strong reliance on a given variable pushes the predictor too strongly in one direction such that other features cannot compensate for this influence on the final score. This SHAP analysis confirms the importance of the “CAD history”, “sex”, and “age” variables as observed in other studies 19 , 20 .

Performance breakdown over different subcohorts and n  = 25 bootstrap draws. The dashed black line indicates the AUROC of a random classifier. Over the full cohort (All Patients), both CARPE Clin. and CARPE ECG reach a statistically significantly higher AUROC than the cardiologist. Additionally, the collaborative approach (CARPE Coll. ) significantly increases predictive performance over CARPE ECG . Please refer to Supplementary Table  6 and Supplementary Fig.  4 for more details. Box plots indicate median (middle line), 25th, and 75th percentile (box). Whiskers extend to points that lie within 1.5 IQRs of the lower and upper quartile. Diamonds are outliers. Error bars in the bar plots indicate 95% confidence intervals. Source data are provided as a Source Data file.

a Bar plots show the mean absolute SHAP value for all clinical variables used by our predictors. Purple scatter plots show individual data points. CAD history and sex are the most important clinical features for both classifiers. The central scatter plots show the impact individual feature values have on the prediction score. High feature values are depicted in a dark blue, low values in a light green. SHAP values for an existing CAD history are always positive. Similarly, SHAP values of the “sex” feature are always positive for male patients. We depict SHAP value distributions over all ages in the scatter plots on the right-hand side. b SHAP values for clinical variables and one 2-6-2 sequence of a patient. The first row shows the feature distribution of the development data set ( n  = 2648) in green. The blue cross marks where in the distribution the patient lies. Second row: SHAP values for the specific patient for each feature over n  = 5 splits. The absence of a CAD history and the resting heart rate of 67 BPM result in negative SHAP values. The patient’s sex (male), his age, and systolic blood pressure at rest are associated with higher SHAP values. Last row: One of the patient’s 2-6-2 sequence (black) with the SHAP values of each individual measurement in the background. We show negative SHAP values in dark purple and positive ones in yellow. Dashed black lines mark the borders of pre-stress, stress, and recovery samples. The largest areas of high SHAP values concentrate in the stress phase around the ST-segment. Error bars in all plots indicate 95% confidence intervals over all models from all five splits. Box plots indicate median (middle line), 25th, and 75th percentile (box). Whiskers extend to points that lie within 1.5 IQRs of the lower and upper quartile. Diamonds are outliers. Bar plots show the mean over n  = 5 test splits with error bars indicating 95% confidence intervals. Source data are provided as a Source Data file.

Overall, the discriminative performance was highest (excluding CARPE Coll. ) in younger patients (CARPE ECG AUROC: 0.78 ± 0.04) in general and in younger patients who did not require pharmacological support specifically (CARPE ECG AUROC: 0.79 ± 0.04). The former cohort also represents the stratum in which the increase over the cardiologist is the highest, namely 0.19 in AUROC and 0.15 in AUPRC. We hypothesise that similar to the conventional ML model (i.e., a random forest), a cardiologist might be more biassed towards a negative diagnosis in younger patients. In contrast, the DL model is more robust to such a behaviour (as shown by the SHAP distribution of the age variable in Fig.  4 ). We show a more detailed assessment of diagnostic performance in different age groups in Fig.  5 .

On the x -axes, we show different age groups in the held-out test and external validation set. Left y -axes: area under the receiver operator characteristic (AUROC). Error bars indicate 95% confidence intervals around the mean. Right y -axes: percentage of patients who comprise the respective subgroup of the x -axis. No cardiologist’s judgement is available in the external validation set, hence CARPE Coll. cannot be evaluated. The performance difference between random forest and CARPE ECG is strongest in the external validation set due to the conventional ML model relying (too) strongly on the “age” variable. Error bars indicate 95% confidence intervals over all models of all five splits. The number of individuals in each bin are 53, 143, 219, 248, 140 for the held-out test set and 281, 341, 208, 86, respectively. Source data are provided as a Source Data file.

On the male subpopulation, CARPE Clin. is outperformed by the cardiologist, indicating that the conventional ML model relies too strongly on the sex feature as an indicator for the presence of fCAD, whereas the DL model and the cardiologist use this feature more effectively. This is underlined by the observation that the performance gap between CARPE ECG and CARPE Clin. is highest in the male subgroup. In female patients, both CARPE Clin. and CARPE ECG perform comparably and better than the cardiologist in terms of AUROC.

In patients of at least 65 years of age, it becomes apparent that human judgement and ML might benefit from each other: while individually, both CARPE Clin. and CARPE ECG perform equally or worse than the cardiologist, combining all predictions in CARPE Coll. results in a statistically significant performance increase over the DL model. It appears that the ML models’ biases are mitigated by the cardiologist’s expertise and vice versa. Augmenting the machine and deep learning output by the cardiologist’s judgement also increases diagnostic performance significantly in the full population. While CARPE Coll. obtains its maximal mean AUROC in the same cohorts as CARPE ECG , the highest mean increase over CARPE ECG can be observed in patients with a CAD history, making it the group in which ML and cardiologists could complement each other most effectively.

Conventional machine learning relies on age, and ST-segment depressions contribute to high risk scores

For the cardiologist who interacts with a risk-stratification tool, it is critical to understand the model’s operations 46 and whether it is consistent with the clinical knowledge about the phenotype. To develop such an understanding, post-hoc explanations 47 can be used to make predictions more interpretable. We use SHAP 38 values, a game-theoretic approach, to explain the outputs of machine learning models. SHAP values provide a score that quantifies the impact an individual feature value has on the model’s prediction. A positive SHAP value is associated with the prediction of the positive class/the presence of fCAD. Conversely, a feature with a negative SHAP value influences the model towards predicting the negative class/the absence of fCAD.

Panel a of Fig.  4 shows mean absolute SHAP values and SHAP value distributions for all clinical variables for CARPE ECG and CARPE Clin. on the left-hand side. On the right-hand side, we show the SHAP values for the “age” feature. For both classifiers, “CAD history” and “sex” are the most influential predictive features (i.e., highest mean absolute value). However, CAD history is only significantly more relevant than the patient’s “sex” in the random forest ( p  = 7.9E-05, test statistic = 7.36 (CARPE Clin. )) and not in CARPE ECG ( p  = 0.055, Welch’s t -test for independent samples, test statistic = 2.24). Furthermore, the SHAP distribution of these variables around the value of zero is strikingly different. While CARPE ECG exhibits many values comparatively close to zero (i.e., there are patients for which the respective features have no significant impact on the model’s final prediction), both CAD history and “sex” have a large impact on the model’s prediction in all patients for the conventional ML model (i.e., the distance to zero for both positive and negative SHAP values is substantial). Additionally, both features show a distinctive separation: each variable instance always leads to either a positive (male and presence of CAD history) or negative (female and absence of CAD history) SHAP value. We observe another distinctively different behaviour in the distribution of SHAP values for the “age” feature. The conventional ML model has learnt an age threshold of 70 years, which, when exceeded, leads to mostly positive SHAP values (i.e., it contributes to predicting the presence of fCAD) and vice versa. CARPE ECG , on the other hand, exhibits a distinctive bell shape around zero, indicating the reduced impact of this variable. While this bias is likely due to the reduced fCAD prevalence of younger patients, the DL model exhibits a more stable behaviour with respect to this variable. The conventional ML model’s reliance on young age as a strong indicator of the absence of fCAD turns out to be detrimental when evaluated on external data, which consists of significantly more young patients (see Fig.  5 ). This underscores the need for explainability and trustworthiness in assessing ML models; if unaddressed, these aspects may preclude clinical applicability.

In addition to performing a population-wide feature relevance analysis, SHAP values allow for sample-specific analyses. In panel b of Fig.  4 , we show a case study of an 83 year-old male patient with no previous CAD. We envision that in a future clinical implementation of our risk assessment tool, such a dashboard will support the cardiologist to understand better on which basis the model arrived at its prediction (e.g., whether the ECG signal is disturbed or noisy) and the influence of each feature (e.g., SHAP values).

The first row of panel b depicts the distributions of the values of all clinical features from the training population. Blue crosses indicate where the patient lies in that distribution. The centre row shows the distribution of SHAP values over five iterations. Moreover, we show the SHAP values of individual measurements in the background of the input ECG in the last row. The mean risk-score CARPE ECG provides for this patient, who was later diagnosed with the presence of fCAD, is 0.77. We show positive SHAP values in yellow, negative ones in dark purple.

Notwithstanding their opposing signs, among the clinical variables, both the absence of a previous CAD and the patient’s age contribute most to the model’s prediction (-0.1 and 0.1, respectively). The normal resting heart rate of 67 is associated with a lower risk score (mean SHAP value: 0.07). While weight, height, and diastolic blood pressure influence the model only marginally, the fact that the patient is male contributes most towards a higher risk score. Similarly, the patient’s age lies above the upper quartile of the training distribution, pushing the model toward predicting a higher score. Lastly, the systolic blood pressure (129 mmHg) also contributes to the prediction of the positive class. The largest contribution that increases the model’s output comes from the ECG. The SHAP values attributed to certain measurements and segments in the ECG might change throughout the different phases of stress testing. In sum, the mean SHAP value for the whole signal is 2.31. The highest SHAP values can be observed in the part of the input signal that comes from the stress phase of the examination. Measurements around the R-peak during rest and, more strikingly, around the ST-segment in the stress and partially in the recovery phase are associated with higher SHAP values than other segments of the ECG. The latter observation is a data-driven and a priori domain-agnostic confirmation of the relevance of ST-segment depression in the diagnosis of fCAD. This is underlined by the fact that in the pre-stress phase, where almost no ST-segment depression is visible, SHAP values around the ST-segment are close to zero. Conversely, negative SHAP values, in line with conventional medical understanding, are observed in the T-wave region during rest, the PR interval during stress, and prominently at the ventricular activation or R-wave peak time. This case study and the relevance of ST-segment depression for the prediction of higher risk scores is supported by a population-wide SHAP analysis in Supplementary Figs.  7 and 8 .

CARPE ECG generalises to unseen data across countries and modalities

To validate our neural network’s generalisation capabilities, we compute its predictive performance on an external validation data set containing 916 consecutive patients referred for exercise myocardial perfusion single photon computed tomography. Referral reasons included non-anginal chest pain, atypical angina, presence of risk factors, or shortness of breath. This data set was retrieved through the THEW data repository 48 (SUI: E-OTH-12-0927-015); it differs from the development data in several key characteristics: First, instead of recording the stress test ECG using bicycle ergometry, it was captured by a treadmill exercise test. Therefore, the resulting signal is subject to noise from walking movements rather than the cycling activity. Second, with a mean age of 55 years, the population in the external data set is significantly younger ( p  = 1.5E-121, one-sided Welch’s t -test, test statistic = 25.39) than the internal study cohort (held-out test set) whose patients are on average 68 years old (see Supplementary Fig.  9 for a complete comparison of all clinical variables). Lastly, the prevalence of ischaemia in the internal cohort is significantly higher compared to the external validation set (7.5%).

As shown in Supplementary Table  7 , both approaches reach a good overall diagnostic performance and perform better on the external data set than on the internal held-out test set. CARPE ECG outperforms the conventional ML model in both AUROC (0.80 \(\pm\) 0.01 vs. 0.75 \(\pm\) 0.004) and AUPRC (0.28 \(\pm\) 0.02 vs. 0.19 \(\pm\) 0.01). We attribute the higher predictive performance of the DL model to the fact that despite coming from a different modality, ECG signals are not fundamentally different among different populations, making it a robust and reliable input signal.

In Fig.  5 , we contrast predictive performance on different age groups in both internal and external validation data. In patients who are younger than 70, both computational approaches consistently outperform the cardiologist in terms of diagnostic accuracy. However, for the stratum that makes up the majority of the data set (ages 70–79), pure computational prediction and human judgement individually perform comparably. However, their combination (CARPE Coll. ) significantly ( p  = 8.1e-4, one-sided Welch’s t -test, test statistic = 7.58) increases diagnostic performance over the cardiologist’s judgement and over CARPE ECG ( p  = 0.001, test statistic = 4.73). The two extremes of the age distribution exemplify how the random forest’s cutoff of 70 years (see SHAP analysis) leads to detrimental performance: The further away a patient group lies from the cutoff, the bigger the performance difference between CARPE ECG and CARPE Clin. becomes. This is even more pronounced in the external validation cohort, where the differences in mean AUROCs (i.e., 10.3 percentage points) are the largest in patients between 26 and 49 years of age.

We derived and validated two ML models for the safe risk-stratification of patients with suspected fCAD. The models were developed using basic clinical information and raw stress test ECG signals. Their performance was compared to a numerical risk estimate by the treating cardiologist after stress testing. Justified by their good predictive performance, we find both models to potentially be useful risk-stratification tools for (a) the primary care setting where cardiac stress testing is not always performed but there is access to the relevant clinical information and (b) the secondary and tertiary setting where stress testing is performed and relevant clinical information is available. Both ensemble learning based on basic clinical information and deep learning based on clinical information and stress ECG signals outperformed the cardiologist’s numerical risk assessment after stress testing as well as currently employed risk scores in predicting the presence of fCAD. At a rule-out probability threshold of <15% as used in current clinical practice guidelines 8 , 21 , compared to the cardiologist, the use of the deep learning model enabled a potential average reduction of myocardial perfusion imaging of 15% without increasing the rate of false negatives at 89% sensitivity and 90% NPV (vs. cardiologist 87% sensitivity and 83% NPV). This was partly due to a consistent risk over- and underestimation at the tails by the cardiologist and a lower diagnostic accuracy in comparison (see Supplementary Fig.  6 ). This shows that calibration is not only a challenge when training ML models but also that the numerical probability estimates from experts must be interpreted with caution. As observed in other studies in ML for healthcare 20 , 22 , 23 , 24 , 49 , we show that conventional ML models based on clinical data alone can be effective predictors and on par with deep learning models when considering the whole data set at once. However, in the context of fCAD prediction, we observe that when stratified into clinically relevant patient subgroups or validated externally, deep learning models consistently yield increased diagnostic performance in most strata. This is likely a direct consequence of phrasing the prediction task as a multi-task problem, thereby preventing overfitting, and the additional data source (ECG signal) allowing the network to learn nuances a conventional ML model cannot capture. For instance, our SHAP analysis revealed that compared to the deep learning model, the ensemble model heavily relied on the “sex” and the “age” feature, with the latter rendering it less generalisable in external validation. A post-hoc interpretability study of the neural network confirmed the relevance of ST-segment depression when predicting fCAD and highlighted the usage of feature attribution methods (such as SHAP values) as potential biomarker discovery tools. In line with previous work 50 , we therefore recommend that any predictive model in cardiology should be assessed internally and externally in terms of (1) its general predictive performance, (2) its effectiveness in clinically relevant subgroups, (3) the relevance of its features, and, if possible, (4) in the context of a human baseline or common clinical practice. In particular, evaluating the degree to which presenting feature attribution values to clinicians may impact their decision-making will be the subject of future prospective studies. Our study showed that combining both computational approaches with the cardiologist’s assessment using logistic regression analysis further increased predictive performance by potentially cancelling out each other’s weaknesses, such as algorithmic or cognitive biases. This combined approach led to a mean AUROC increase of four percentage points over the DL model in patients with a CAD history and an increase of 17 percentage points over the cardiologist numerical prediction in patients below the age of 65 who possessed the capacity to undergo the stress test unaided by pharmacological support. Several limitations should be considered when interpreting our findings. Although we have used a stringent methodology to adjudicate the presence or absence of fCAD, we still may have misclassified a small number of patients. Reflecting clinical practice, the expert interpretation of fCAD was not blinded to clinical and stress ECG data, which could have led to an overestimation of these features. Nevertheless, the model’s discriminative performance remained consistent across patients, regardless of whether they received invasive coronary artery assessment or not (Supplementary Figs.  2 and 3 ), indicating a minimal impact of this non-blinded approach. The model was developed in symptomatic patients referred to a tertiary hospital. During study enrolment, MPI-SPECT/CT was the standard non-invasive imaging modality and was applied to patients with a wide range of pre-test probability for CAD. Based on the pre-test probabilities employed in the current ESC guidelines for the diagnosis and management of chronic coronary syndromes 21 , 29% of the patients included in the derivation and internal validation cohort would be classified as low risk (<15% probability). As in most other cohorts enroling consecutive patients with suspected CAD, women were underrepresented in the overall cohort. Accordingly, some of the subgroup analyses may have been underpowered in female patients. Similarly, patients of African or Asian descent were underrepresented in this study, and potential differences between these groups cannot be addressed. In the derivation cohort, 26% of patients were below 60, and 7.6% were below the age of 50. Therefore, the results of this study might not apply to very young patients. While the value of more advanced neural network architectures (e.g., attention-based) and ensemble methods (e.g., gradient-boosted trees) may also be explored in the future, prospective clinical studies must be prioritised to establish the clinical value of the CARPE models, interpretability dashboards, and collaborative machine learning.

We integrated the clinical judgement of physicians into our machine learning model using logistic regression, which further increased its accuracy. However, it is important to acknowledge the limitations of logistic regression analysis and that in real-world clinical practice physicians are unaware of the influence their numerical predictions have on the model’s score. The extent to which this knowledge gap influences subsequent risk assessments and thus the model’s performance in real-world clinical practice remains an open question. Moreover, clinical utility and the generalisation capabilities of our method are affected by distribution shifts in the input data. In particular, the dependency of clinical judgement on the practitioner’s level of experience and their intuitive understanding of the patient’s medical history warrants a recalibration of both the ensemble model and the logistic regression before deployment in novel clinical environments. Thus, the observed improvement in performance through logistic regression analysis may not directly reflect the clinical utility or practical applicability of its predictions in healthcare settings. For our collaborative approach to achieve this, careful model recalibration, score interpretation, as well as continuous monitoring of clinical outcomes will be required. As a leading cause of mortality and morbidity worldwide, fCAD is affecting an ever-increasing patient population. With the concurrent demographic ageing in most high- and middle-income countries, there is a major clinical need for safe, accessible, effective, and cost-efficient risk stratification tools to identify patients. Our research underscores the potential clinical utility of ML in reducing potentially unwarranted examinations to support clinicians in providing the best possible care for their patients. Ultimately, maximising predictive performance and clinical acceptance most likely necessitates integrating human judgement with ML predictions in some way.

The BASEL VIII study was approved by the local ethics committee (swissethics, BASEC, Ethikkommission Nordwest- und Zentralschweiz) under the number EKBB 100/04 and carried out according to the principles of the Declaration of Helsinki.

Study population

This analysis is part of a large prospective diagnostic study (NCT01838148, clinicaltrials.gov) designed to advance the early detection of inducible myocardial ischaemia 51 , 52 . Consecutive adult patients referred to the University Hospital Basel, Switzerland for rest/stress myocardial perfusion single-photon emission tomography/computer tomography (MPI-SPECT/CT) with symptoms possibly due to inducible myocardial ischaemia were enrolled between January 2010 to May 2016. During that period MPI-SPECT/CT was the preferred imaging modality in patients with a wide range of pre-test probabilities for functionally relevant CAD 29 , 52 . All patients provided written informed consent. Participants did not receive any form of financial compensation or equivalent benefits for their participation in this study. Clinical information, including patient characteristics, medications, symptoms, and prior cardiovascular history were documented by physicians using standardised questionnaires and all medical files available. Based on all clinical information prior to testing, the treating physician recorded a subjective clinical assessment regarding the presence of inducible myocardial ischaemia due to CAD on a visual analogue scale with values between 0% and 100%. Supplementary Table  9 shows demographic and clinical characteristics of patients in development and held-out test set. The sex of the patient was based on the medical files of the University Hospital Basel which represent the Swiss civil status register (i.e., “Geschlecht/Sex” as listed on the passport). We only disaggregated data by sex not by gender, as the latter has not been collected.

ECG preprocessing and feature extraction

To prepare the 12-lead ECG signals as input to our deep learning approach, we first performed a small number of signal preprocessing steps including downsampling, smoothing, and outlier removal. However, as the choice of preprocessing can affect the ECG’s morphology 53 , preprocessing parameters were determined by a grid search (see Supplementary Fig.  5 ). Simultaneously, we evaluated the predictive performance of individuals leads and their combination and selected the best performing combination for evaluation on the held-out test set. Secondly, scalability limitations imposed by the neural network architecture require a significant reduction of the length of the ECG input signal from ~500,000 time points (i.e. 15 min) to 5000. For this, we sample 2 s from the beginning of the examination, 6 s from the last 2 min of the stress phase, and 2 s from the last 3 min of the recovery phase and merge them into a single time series (see panel b in Fig.  1 ) whose information content is dominated by the stress phase. This time series, which we refer to as the 2-6-2 sequence, was constructed up to twenty times per patient by sampling the subsequences at different time points. Such summary sequences represent a compromise between expressivity (each sequence contains information from the warm-up, stress, and recovery phase) and computational scalability.

Exercise stress testing protocol and ECG raw data acquisition

Resting heart rate, blood pressure, and 12-lead resting ECG were recorded before exercise. A standardised, stepwise, and symptom-limited upright bicycle exercise test was performed 54 , 55 . Beta-blockers and antianginal medication were paused for at least 48 h and nitrates for at least 24 h before testing. Exercise stress testing was considered conclusive if 85% of the age predicted maximum heart rate was reached. If this was not feasible, physical exercise was stopped and patients were switched to an adenosine or dobutamine pharmacologic stress testing protocol 51 , 55 , 56 . In patients in whom physical stress testing was contraindicated or the target exercise performance was not reached, pharmacological testing was performed. After testing and blinded to the MPI-SPECT/CT results, the treating physician once more recorded a clinical post-test probability regarding the presence of fCAD on a visual analogue scale (0–100%). The 12-lead ECG signals were recorded with two different devices (Schiller AT-110 and Schiller CS-200 Excellence) at 500 Hz and 1000 Hz with a minimal resolution of 5 μV/bit and a minimal diagnostic signal bandwidth of 0.05 Hz to 150 Hz.

Adjudication of fCAD

Adjudication of fCAD was based on expert interpretation of MPI-SPECT/CT images combined with information obtained from invasive coronary angiography and fractional flow reserve measurements, whenever available. All patients underwent a routine standard rest/stress dual isotope ( 201 Tl for rest, 99m Tc sestamibi for stress) or a single isotope ( 99m Tc sestamibi for stress and rest) MPI-SPECT/CT protocol. MPI-SPECT/CT images were scored semi-quantitatively using a 17-segment model with a 5-point scale (0 = normal, 1 = mildly reduced tracer uptake, 2 = moderately reduced uptake, 3 = severely reduced uptake and 4 = no uptake). Summed stress score and summed rest score were calculated by adding the scores of the 17 segments in the stress and rest images. Summed difference score was the difference between summed stress score and summed rest score. A summed difference score of at least 2 or positive transient ischaemic dilation ratio (≥1.22 for the dual isotope protocol and 1.12 for the single isotope protocol) was considered as fCAD 27 , 28 , 29 . Summed stress score and summed rest score were derived by visual assessment of two expert readers and compared with the software result. Differences in the visual assessment were resolved by finding consensus. In case of equivocal findings from MPI-SPECT/CT and coronary angiography, an adjudication committee of two independent cardiologists (one interventional cardiologist, one general cardiologist) that were blinded to study biomarker results reviewed the case using all clinically available data. A positive perfusion scan was overruled when coronary angiography showed normal coronary arteries, while a negative perfusion scan was overruled if coronary angiography (within 3 months) either revealed a high-grade coronary lesion (>75%) or if there was fractional flow reserve below 0.80. In total the adjudication committee reviewed 147 cases or 21% of the 701 patients that underwent coronary angiography within 90 days.

Neural network architecture

Supplementary Fig.  10 provides an overview of the multi-task learning neural network architecture. While the patient’s static data X clin is embedded by a neural network, the ECG data (X ecg ) is fed into a residual neural network akin to the one used by Ribeiro et al. 15 . The concatenation of the output of the embedding layer for the clinical variables and the output of the residual network serves as input to four subnetworks, each of which is responsible for the prediction of one of the four tasks. Each task has its own loss function. More details about the exact layer definitions can be found in Supplementary Fig.  3 .

ST-segment depression

The human baseline is complemented by an automated determination of ST-segment depression. While this morphological feature is commonly linked to ischaemia 39 , 40 , the exact time points in the ECG at which ST-amplitude is measured varies 4 . We compute ST-segment depression as follows. First, we perform a QRS-delineation using the neurokit2 software package 4 , 57 on the complete stress test ECG. Then, we determine the mean isoelectric line for each stress phase of a given 2-6-2 sequence. For this, we take the mean of the last l PR milliseconds preceding the Q-wave over all heartbeats in a specific stress phase. Similarly, we determine the mean ST-amplitude for each 2-6-2 stress phase by using the ECG measurement 60 ms after the J-point. The mean ST-segment depression (difference between mean isoelectric line and ST-amplitude) is determined for each stress phase (ST Pre , ST Stress , ST Rec ). The differences between ST Stress /ST Rec and baseline ST-depression (ST Pre ) are then aggregated over all 2-6-2 sequences of a patient by using either their mean, median, minimum, or maximum. Importantly, the physiological response to stress may differ among the patients subject to different stress types. Therefore, the parameter grid shown in Supplementary Table  10 is evaluated separately for all three cohorts and all leads.

Data splits and bootstrapping

We split the data set 3:1 into a development and held-out test set containing 2648 and 874 patients, respectively (see Supplementary Fig.  11 ). During the development of the model, we had no access to the held-out test set. Access was provided once we fixed all model parameters. The development set was further divided into 5 stratified splits of training, validation, and calibration set, where the latter makes up 10% of the training set. The ratio of training to validation set size is 4:1. Each of five splits of the development set contained on average 36977, 9254, and 5129 train, validation, and calibration 2-6-2 sequences from 1882, 471, and 260 patients, respectively. If not specified otherwise, bootstrapping has been performed to obtain distributions for statistical testing. This was done by pooling all predictions from all five splits and sampling 80% in 25 different draws. Since the cardiologist only scores each patient once, we sampled the same patients for the cardiologist that have been selected for the computational methods in each draw. This way each draw contains predictions for the same patients from the different predictors to be compared.

Statistics and reproducibility

All p -values for the comparison of performance metrics (i.e., AUROC, AUPRC) were computed on the distributions obtained by the bootstrapping procedure described in the previous section. For this, a one-sided Kolmogorov-Smirnov test was used. Multiple hypotheses are corrected for using Bonferroni correction. When comparing the age distribution of our data set with the external data set, a one-sided Welch’s t -test was used. To compare the odds of obtaining a positive fCAD label with/without a history of CAD, we used a two-sided Fisher’s exact test. The comparison of SHAP values was performed over the distributions of five data splits as described above; we used Welch’s t -test for independent samples. The development set was split into training/validation/calibration in a stratified manner ( n  = 5), meaning the fCAD prevalence remains the same in all splits. From the original data set, 697 patients were excluded because no digital ECG data was available (see Supplementary Fig.  11 ). Reflecting clinical practice, the expert interpretation of fCAD was not blinded to clinical and stress ECG data. However, the treating physician was blinded to the MPI-SPECT/CT results when submitting the post-test probability score. With access to the THEW data set (SUI: E-OTH-12-0927-015), all results pertaining to this data set can be reproduced using the publicly available code at https://github.com/BorgwardtLab/CARPE 58 .

Preprocessing, lead selection, and auxiliary task regularisation

All 1000 Hz signals were downsampled to 500 Hz. ECG signals from exercise stress testing are subject to high noise levels from various sources. To assess the influence of noise on classification performance, we consider the following preprocessing schemes: 1. No preprocessing, 2. minimal preprocessing with a high-pass Butterworth filter of order five, and a cutoff frequency of 0.5 Hz followed by moving average smoothing, and 3. a thorough preprocessing pipeline consisting of a wider bandpass filter (0.05 Hz–150 Hz), moving-median subtraction to remove baseline wandering, a Savitzky–Golay filter 59 for smoothing, and winsorizing to deal with spurious outliers.

To evaluate the impact that individual ECG leads, preprocessing, and auxiliary tasks have on predictive performance, we proceeded as follows: First, we used the first development split to determine the most promising leads (in terms of area under the precision-recall curve (AUPRC) on the validation set) by performing a grid search over (a) three preprocessing schemes described above, and (b) learning rate parameters \(\eta \in \{0.01,\,0.001,\,0.001\}\) for all twelve leads individually and in combination. More specifically, we trained \(13\,\times 3\,\times 3=117\) neural networks to determine the three best performing leads. The first number accounts for the 12 individual ECG leads plus one configuration that combines all leads. The second number represents three preprocessing schemes and is followed by the number of learning rates that were analysed. Subsequently, we picked the three best-performing leads and their respective preprocessing/learning rate combination to assess the impact of all auxiliary tasks. In order to do so, the performance on the validation set was averaged over all splits on a \(5\,\times 5\,\times 5\,\times 3\) parameter grid as shown in Supplementary Table  4 . Finally, the best-performing model was enriched with clinical variables to receive the final model, which we evaluated on the held-out test set. The results of this analysis on the validation set of the development data set are shown in Supplementary Fig.  5 and Supplementary Table  5 .

Calibration

Supplementary Fig.  6 depicts the calibration of CARPE ECG , CARPE Clin. , and cardiologist on both training and held-out test data. The red dashed lines indicate the two decision cutoffs (5% and 15%) as advocated in European and US-American guidelines 8 , 21 . On the training set, CARPE ECG is almost perfectly calibrated at 5% but slightly overestimates fCAD probability at 15%. On the held-out test set, CARPE ECG remains close to the diagonal but now underestimates the presence of fCAD. The cardiologist underestimates the presence of fCAD at both thresholds and in both data sets, yet performs similarly to CARPE ECG at the 15% on the held-out test set. The ensemble method significantly overestimates the presence of fCAD around the relevant decision thresholds on the training set and reaches best calibration on the held-out test set at the 5% cutoff. At the 15% threshold, however, it continues to overestimate the presence of fCAD. Both computational methods exhibit a significantly lower Brier score (hence are better calibrated than the cardiologist) on the held-out test set. More precisely, the cardiologist reaches a score of 0.23 ± 0.009, the random forest 0.22 ± 0.004 ( p  = 3.59E-5, one-sided t -test, df = 24, test statistic = -4.78), and CARPE ECG 0.18 ± 0.006 ( p  = 7.97E-16, one-sided t -test, df = 24, test statistic = -18.15).

External validation data

The external dataset consists of 927 consecutive patients referred for exercise myocardial perfusion single photon emission computed tomography (SPECT) at Assuta Medical Center and Sheba Medical Center and is a subset of the data presented by Sharir et al. 49 Throughout the baseline, exercise, and recovery phases, a high-resolution 12-lead ECG was continuously recorded using the HyperQ Stress System from BSP Ltd, Tel Aviv, Israel, at a rate of 1000 samples/second with 16-bit resolution and an analogue frequency response of 0.05 to 125 Hz (measurement sensitivity <0.15 V). Patients with a cardiac pacemaker, atrial fibrillation at the time of testing, or a QRS duration equal to or greater than 120 milliseconds were excluded from the study. The exercise procedure was carried out as follows: beta blockers and calcium channel blockers were discontinued at least 48 h prior to the test, and a symptom-limited treadmill exercise test was conducted using the Bruce protocol.

In contrast to our internal training and held-out test set, the external data set is not annotated with the stress phases (pre, stress, recovery). We therefore compute the heart rate for each minute of the ECG signal and use its maximum as the end of the stress phase. This allows us to extract 2-6-2 sequences that are equivalent to the training data. Additionally, we exclude patients for which the required clinical variables are missing. If the value for a single variable is missing, we exclude the patient. This is the case for the ground truth label (missing in nine patients), the weight variable (missing in one patient), and the height variable (missing in one patient). Supplementary Fig.  9 depicts the distribution of relevant clinical variables of the internal and external data sets. The biggest difference between both data sets can be observed in the age variables (patients from the internal data set are significantly older) and the fCAD/ischaemia prevalence.

Additional SHAP analysis

To measure the impact ECG segments have on the prediction, we summarize their SHAP values by summing them to express a segment’s contribution in a single scalar value. We perform this aggregation for all ECG segments in all 2-6-2 sequences. We then stratify each segment by its stress phase to investigate whether the origin of the segment (in terms of the stress test phase) influences the attributed contribution. Supplementary Fig.  7 shows the result of this analysis in patients for which a low and high CAD probability was predicted. Overall, we observe more segments with SHAP values deviating from zero in the higher-risk population. In particular, the ST-Segment in the stress phase is associated with particularly high SHAP values. For the population for which a low risk was predicted, the SHAP values from the QRS complex from the stress phase contribute, on average, most to the prediction signal. We summarize both the QRS complex and the ST-segment from the respective cohorts in Supplementary Fig.  8 .

To investigate whether specific ECG patterns exist that contribute to a lower/higher predicted CAD score, we perform the following analysis: If a stress test phase (i.e., Pre/Stress/Recovery) of a 2-6-2 sequence contains an ECG pattern (e.g., P-wave, ST-segment) whose summed SHAP score exceeds/is lower than a threshold, we extract all ECG waves (i.e., from P-wave onset to T-wave offset) from this phase. Each extracted ECG wave contains a representative of the pattern that has a high influence on the model’s prediction. Motivated by the results visualized in Supplementary Fig.  7 , we choose ECG waves whose QRS-complex has a SHAP score of maximal -0.25 to highlight a pattern contributing to the prediction of the absence of CAD. Similarly, we selected ECG waves whose ST-segment has a SHAP score of at least 0.25 as a pattern that contributes to the prediction of the presence of CAD. Furthermore, we limit this analysis to samples with higher predictive score and apply the same probability thresholds (15% and 85%) as in the previous analysis. We align all waves using dynamic time warping and visualize both the aligned waves (grey) and their mean wave (red) in Supplementary Fig.  8 .

Calibration and performance for patients with and without ICA

Supplementary Figs.  2 and 3 show calibration and predictive performance (AUROC and AUPRC) for patients who did and did not undergo ICA within 90 days. The respective prevalence of fCAD in the held-out test set was 84% and 18%, respectively. In the latter cohort, the shown predictors are calibrated similarly to the full cohort (viz. Supplementary Fig.  6 ). In the subgroup of patients who underwent ICA within 90 days (i.e., overall high-risk patients), we observe that all predictors, including the cardiologist, underestimate the presence of fCAD. Considering AUROC, CARPE ECG exhibits a better mean predictive performance in both cohorts. In patients with ICA, the mean performance gain is 18 percentage points (CARPE ECG : 0.72 ± 0.06, Cardiologist: 0.54 ± 0.06) and 11 percentage points (CARPE ECG : 0.72 ± 0.02, Cardiologist: 0.61 ± 0.03). Furthermore, in patients who did not undergo ICA within 90 days, the combination of cardiologist and deep learning method, CARPE Coll. modestly increases mean predictive performance from 0.72 to 0.74.

Statistical interaction tests by subgroups

Supplementary Fig.  4 provides a statistical interaction analysis based on the performance increase of CARPE ECG over the cardiologist. On the complete held-out test cohort, this increase is, on average, 8 percentage points. The only variable showing a statistically significant ( p  = 0.0067, two-sided Z -test) effect is “Age,” where the performance difference between CARPE ECG and the cardiologist is significantly higher in patients <65 years of age compared to patients who are at least 65 years old.

Data collection and data analysis software

The 12-lead ECG signals were recorded with two ECG machines, namely a Schiller AT-110 and Schiller CS-200 Excellence. Exported.ful and.xml files were analysed with the standard xml module of python 3.8 and numpy at version 1.24.4. ECG preprocessing was performed using the ‘signal’ module of scipy at version 1.10.1. QRS Delineation was performed using the matlab ecg-kit version 1.4.0.0. To train our models and analyse and visualize the data, we used the following python libraries and versions ipython version 8.12.3, jupyter version 1.0.0, matplotlib version 3.7.3, networkx version 2.8.8, notebook version 7.0.4, pandas version 2.0.3, pytorch-lightning version 1.2.1, scikit-learn version 1.2.2, scipy version 1.10.1 seaborn version 0.13.0, pytorch version 1.6.0, torchmtl version 0.1.9.

Reporting summary

Further information on research design is available in the  Nature Portfolio Reporting Summary linked to this article.

Data availability

The data that support some of the findings of this study are not openly available due to reasons of sensitivity of patient data and are available from the corresponding author ([email protected]) upon request. The request should include the name and full contact information of the person and institution requesting the data, the specific identification of the data being requested and the purpose of requesting the data. Data requests under agreement will be considered for purposes of reproducing the data presented herein, subject to appropriate confidentiality obligations and restrictions. The timeframe for response to requests is estimated to be four to 8 weeks and restrictions imposed on data use will be individualized by case-by-case data use agreements. The data resides in the secured IT infrastructure of the University Hospital Basel and respective files can be shared after anonymization upon individual request. Data used for external validation was provided by the Telemetric and Holter ECG Warehouse of the University of Rochester (THEW), NY. It cannot be made public by the authors. To obtain access, interested parties must register with the THEW project ( http://thew-project.org/ ), submit a research proposal, and fill out the data usage agreement for the dataset with identifier E-OTH-12-0927-015. For-profit organisations may also purchase the data set for an access fee as detailed on the website. The authors declare that all data supporting the findings of this study which are not protected by patient privacy considerations, are available within the paper, its supplementary information files and downloadable files deposited at figshare ( https://doi.org/10.6084/m9.figshare.25514644 ).

Code availability

Preprocessing scripts, trained neural network model checkpoints and random forest classifier are publicly available at https://github.com/BorgwardtLab/CARPE 58 .

Townsend, N. et al. Cardiovascular disease in europe: epidemiological update 2016. Eur. Heart J. 37 , 3232–3245 (2016).

Article   PubMed   Google Scholar  

Writing Group Members et al. Heart disease and stroke statistics-2016 update: a report from the American Heart Association. Circulation 133 , e38–e360 (2016).

Google Scholar  

GBD 2017 Disease and Injury Incidence and Prevalence Collaborators. Global, regional, and national incidence, prevalence, and years lived with disability for 354 diseases and injuries for 195 countries and territories, 1990–2017: a systematic analysis for the global burden of disease study 2017. Lancet 392 , 1789–1858 (2018).

Puelacher, C. et al. Diagnostic value of ST-segment deviations during cardiac exercise stress testing: systematic comparison of different ECG leads and time-points. Int. J. Cardiol. 238 , 166–172 (2017).

Ladapo, J. A., Blecker, S. & Douglas, P. S. Physician decision making and trends in the use of cardiac stress testing in the United States: an analysis of repeated cross-sectional data. Ann. Intern. Med. 161 , 482–490 (2014).

Article   PubMed   PubMed Central   Google Scholar  

Rozanski, A. et al. Temporal trends in the frequency of inducible myocardial ischemia during cardiac stress testing: 1991 to 2009. J. Am. Coll. Cardiol. 61 , 1054–1065 (2013).

Devereaux, P. J. The potential for troponin to inform prognosis in patients with stable coronary artery disease. Ann. Intern. Med. 169 , 808–809 (2018).

Article   CAS   PubMed   Google Scholar  

Knuuti, J. et al. 2019 ESC Guidelines for the diagnosis and management of chronic coronary syndromes. Eur. Heart J. 41 , 407–477 (2020).

Juarez-Orozco, L. E. et al. Impact of a decreasing pre-test probability on the performance of diagnostic tests for coronary artery disease. Eur. Heart J. Cardiovasc. Imaging 20 , 1198–1207 (2019).

Ansari, S. et al. A review of automated methods for detection of myocardial ischemia and infarction using electrocardiogram and electronic health records. IEEE Rev. Biomed. Eng. 10 , 264–298 (2017).

Johnson, K. W. et al. Artificial intelligence in cardiology. J. Am. Coll. Cardiol. 71 , 2668–2679 (2018).

Bizopoulos, P. & Koutsouris, D. Deep learning in cardiology. IEEE Rev. Biomed. Eng. 12 , 168–193 (2019).

Strodthoff, N., Wagner, P., Schaeffter, T. & Samek, W. Deep learning for ECG analysis: benchmarks and insights from PTB-XL. IEEE J. Biomed. Health Inf. 25 , 1519–1528 (2021).

Article   Google Scholar  

Hannun, A. Y. et al. Cardiologist-level arrhythmia detection and classification in ambulatory electrocardiograms using a deep neural network. Nat. Med. 25 , 65–69 (2019).

Article   CAS   PubMed   PubMed Central   Google Scholar  

Ribeiro, A. H. et al. Automatic diagnosis of the 12-lead ECG using a deep neural network. Nat. Commun. 11 , 1760 (2020).

Overmars, L. M. et al. Preventing unnecessary imaging in patients suspect of coronary artery disease through machine learning of electronic health records. Eur. Heart J. - Digital Health 3 , 11–19 (2022).

Kukar, M., Kononenko, I., Groselj, C., Kralj, K. & Fettich, J. Analysing and improving the diagnosis of ischaemic heart disease with machine learning. Artif. Intell. Med. 16 , 25–50 (1999).

Alizadehsani, R. et al. Machine learning-based coronary artery disease diagnosis: a comprehensive review. Comput. Biol. Med. 111 , 103346 (2019).

Megna, R. et al. A comparison among different machine learning pretest aproaches to predict stress-induced ischemia at PET/CT myocardial perfusion imaging. Comput. Math. Methods Med. 2021 , 3551756 (2021).

Miller, R. J. H. et al. Machine learning to predict abnormal myocardial perfusion from pre-test features. J. Nucl. Cardiol. 29 , 2393–2403 (2022).

Gulati, M. et al. 2021 AHA/ACC/ASE/CHEST/SAEM/SCCT/SCMR guideline for the evaluation and diagnosis of chest pain: a report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation 144 , e368–e454 (2021).

PubMed   Google Scholar  

Chen, D. et al. Deep learning and alternative learning strategies for retrospective real-world clinical data. NPJ Digit Med. 2 , 43 (2019).

Hyland, S. L. et al. Early prediction of circulatory failure in the intensive care unit using machine learning. Nat. Med. 26 , 364–373 (2020).

Lewis, M. et al. Comparison of deep learning with traditional models to predict preventable acute care use and spending among heart failure patients. Sci. Rep. 11 , 1164 (2021).

Article   ADS   CAS   PubMed   PubMed Central   Google Scholar  

Verberne, H. J. et al. EANM procedural guidelines for radionuclide myocardial perfusion imaging with SPECT and SPECT/CT: 2015 revision. Eur. J. Nucl. Med. Mol. Imaging 42 , 1929–1940 (2015).

Sou, S. M. et al. Direct comparison of cardiac troponin I and cardiac troponin T in the detection of exercise-induced myocardial ischemia. Clin. Biochem. 49 , 421–432 (2016).

Lee, G. et al. Clinical benefit of high-sensitivity cardiac troponin I in the detection of exercise-induced myocardial ischemia. Am. Heart J. 173 , 8–17 (2016).

Tanglay, Y. et al. Incremental value of a single high-sensitivity cardiac troponin I measurement to rule out myocardial ischemia. Am. J. Med. 128 , 638–646 (2015).

Walter, J. E. et al. Prospective validation of a biomarker-based rule out strategy for functionally relevant coronary artery disease. Clin. Chem. 64 , 386–395 (2018).

Otles, E. et al. Mind the performance gap: examining dataset shift during prospective validation. In Proc. 6th Machine Learning for Healthcare Conference Vol. 149 (eds. Jung, K., Yeung, S., Sendak, M., Sjoding, M. & Ranganath, R.) 506–534 (PMLR, 2021).

Breiman, L. Random forests. Mach. Learn. 45 , 5–32 (2001).

Cortes, C. & Vapnik, V. Support-vector networks. Mach. Learn. 20 , 273–297 (1995).

Caruana, R. Multitask Learning. Machine Learning 28 , 41–75 (1997).

He, K., Zhang, X., Ren, S. & Sun, J. Deep residual learning for image recognition. In Proc. IEEE Conference on Computer Vision and Pattern Recognition https://doi.org/10.1109/CVPR.2016.90 (IEEE, Las Vegas, USA, 2016).

Bock, C. torchMTL: A Lightweight Module For Multi-Task Learning In Pytorch. https://github.com/chrisby/torchMTL (2020) https://doi.org/10.5281/zenodo.4362515 .

Ruder, S. An overview of multi-task learning in deep neural networks. https://doi.org/10.48550/arXiv.1706.05098 (2017).

Shapley, L. S. 17. A value for n-person games. In Contributions To The Theory of Games (AM-28), Volume II (eds. Kuhn, H. W., Tucker, A. W.) 307–318. https://www.degruyter.com/document/doi/10.1515/9781400881970-018/html (Princeton University Press, 2016).

Lundberg, S. M. & Lee, S.-I. A unified approach to interpreting model predictions. In Advances in Neural Information Processing Systems , (eds. Guyon, I. et al.) Vol 30. https://proceedings.neurips.cc/paper_files/paper/2017/file/8a20a8621978632d76c43dfd28b67767-Paper.pdf (Curran Associates, Inc., 2017).

Stern, S. State of the art in stress testing and ischaemia monitoring. Card. Electrophysiol. Rev. 6 , 204–208 (2002).

Pollehn, T. The electrocardiographic differential diagnosis of ST segment depression. Emerg. Med. J. 19 , 129–135 (2002).

Vickers, A. J. & Elkin, E. B. Decision curve analysis: a novel method for evaluating prediction models. Med. Decis. Mak. 26 , 565–574 (2006).

Vickers, A. J., van Calster, B. & Steyerberg, E. W. A simple, step-by-step guide to interpreting decision curve analysis. Diagn. Progn. Res. 3 , 18 (2019).

Vickers, A. J., Van Calster, B. & Steyerberg, E. W. Net benefit approaches to the evaluation of prediction models, molecular markers, and diagnostic tests. BMJ 352 , https://doi.org/10.1136/bmj.i6 (2016).

Schlesinger, D. E. & Stultz, C. M. Deep learning for cardiovascular risk stratification. Curr. Treat. Options Cardiovasc. Med. 22 , 15 (2020).

Genders, T. S. S. et al. A clinical prediction rule for the diagnosis of coronary artery disease: validation, updating, and extension. Eur. Heart J. 32 , 1316–1330 (2011).

Biran, O. & Cotton, C. Explanation and Justification in Machine Learning: A Survey . http://www.cs.columbia.edu/~orb/papers/xai_survey_paper_2017.pdf (2017).

Lipton, Z. C. The mythos of model interpretability. Queue 16 , 31–57 (2018).

Couderc, J.-P. The telemetric and Holter ECG warehouse initiative (THEW): a data repository for the design, implementation and validation of ECG-related technologies. Conf. Proc. IEEE Eng. Med. Biol. Soc. 2010 , 6252–6255 (2010).

PubMed Central   Google Scholar  

Sharir, T. et al. Use of electrocardiographic depolarization abnormalities for detection of stress-induced ischemia as defined by myocardial perfusion imaging. Am. J. Cardiol. 109 , 642–650 (2012).

Ghassemi, M., Oakden-Rayner, L. & Beam, A. L. The false hope of current approaches to explainable artificial intelligence in health care. Lancet Digit Health 3 , e745–e750 (2021).

Walter, J. et al. Using high-sensitivity cardiac troponin for the exclusion of inducible myocardial ischemia in symptomatic patients: a cohort study. Ann. Intern. Med. 172 , 175–185 (2020).

Mueller, D. et al. Direct comparison of cardiac troponin T and I using a uniform and a sex-specific approach in the detection of functionally relevant coronary artery disease. Clin. Chem. 64 , 1596–1606 (2018).

Article   ADS   PubMed   Google Scholar  

Buendía-Fuentes, F. et al. High-bandpass filters in electrocardiography: source of error in the interpretation of the ST segment. ISRN Cardiol. 2012 , 706217 (2012).

Bourque, J. M. & Beller, G. A. Value of exercise ECG for risk stratification in suspected or known CAD in the era of advanced imaging technologies. JACC Cardiovasc. Imaging 8 , 1309–1321 (2015).

Schaerli, N. et al. Incremental value of high-frequency QRS analysis for diagnosis and prognosis in suspected exercise-induced myocardial ischaemia. Eur. Heart J. Acute Cardiovasc Care 9 , 836–847 (2020).

Wagener, M. et al. Diagnostic and prognostic value of lead aVR during exercise testing in patients suspected of having myocardial ischemia. Am. J. Cardiol. 119 , 959–966 (2017).

Makowski, D. et al. NeuroKit2: A Python toolbox for neurophysiological signal processing. Behav. Res. Methods 53 , 1689–1696 (2021).

C. Bock. et al. Enhancing The Diagnosis Of Functionally Relevant Coronary Artery Disease With Machine Learning. https://github.com/BorgwardtLab/CARPE (2024) https://doi.org/10.5281/ZENODO.10868173 .

Savitzky, A. & Golay, M. J. E. Smoothing and differentiation of data by simplified least squares procedures. Anal. Chem. 36 , 1627–1639 (1964).

Download references

Acknowledgements

This study was supported by the Alfried Krupp Prize for Young University Teachers of the Alfried Krupp von Bohlen und Halbach-Stiftung (K.B.). The founders had no influence on the study question or design in any way. Data used for this research was provided by the Telemetric and Holter ECG Warehouse of the University of Rochester (THEW), NY.

Open Access funding enabled and organized by Projekt DEAL.

Author information

These authors contributed equally: Christian Bock, Joan Elias Walter, Bastian Rieck.

These authors jointly supervised this work: Karsten Borgwardt, Christian Müller.

Authors and Affiliations

Department of Biosystems Science and Engineering, ETH Zürich, Basel, Switzerland

Christian Bock, Bastian Rieck & Karsten Borgwardt

Swiss Institute for Bioinformatics, Lausanne, Switzerland

Cardiovascular Research Institute Basel, University Hospital of Basel, University of Basel, Basel, Switzerland

Joan Elias Walter, Ivo Strebel, Klara Rumora, Ibrahim Schaefer, Michael J. Zellweger & Christian Müller

Department of Cardiology, University Hospital of Basel, University of Basel, Basel, Switzerland

Department of Endocrinology, Diabetology and Clinical Nutrition, University Hospital Zurich, University of Zurich, Zurich, Switzerland

Joan Elias Walter

Institute of AI for Health, Helmholtz Munich and Technical University of Munich, Munich, Germany

Bastian Rieck

Department of Machine Learning and Systems Biology, Max Planck Institute of Biochemistry, Martinsried, Germany

Karsten Borgwardt

You can also search for this author in PubMed   Google Scholar

Contributions

C.B., J.E.W., B.R., I.St., K.B., C.M. designed the experiments; J.E.W, I.St., K.R., I.Sc., M.J.Z., C.M. collected and provided both clinical and ECG data; C.B. and I.St. preprocessed the raw ECG signals with contributions from B.R. and J.E.W.; C.B. and B.R. developed and implemented the machine learning pipelines; C.B. performed all experiments with contributions from B.R., J.E.W., I.St., K.B., C.M.; J.E.W. and C.M. performed the clinical interpretation of results and provided C.B., B.R., K.B. with relevant clinical context. C.B. created all figures with support from B.R., J.E.W., K.B.; C.B., B.R., C.B. and J.E.W. performed statistical analyses with contributions from K.B.; K.B., C.M., J.E.W. conceived and directed the project; C.B., J.E.W., B.R., C.M., K.B. wrote the manuscript with the assistance of feedback of all the other co-authors.

Corresponding authors

Correspondence to Karsten Borgwardt or Christian Müller .

Ethics declarations

Competing interests.

J.E.W. has no conflict of interest to declare regarding this project and reports grants from Swiss Heart Foundation (FF19097 and F18111) and from the Swiss Academy Medical Sciences. C.M. has no conflict of interest to declare regarding this project and received research support from the Swiss National Science Foundation, the Swiss Heart Foundation, the KTI, the University of Basel, the University Hospital Basel, Abbott, Beckman Coulter, Brahms, Idorsia, Novartis, Ortho Clinical Diagnostics, Quidel, Roche, Siemens, Singulex, and Sphingotec as well as speaker honoraria/consulting honoraria from Amgen, AstraZeneca, Bayer, Beckman Coulter, Boehringer Ingelheim, BMS, Idorsia, Novartis, Osler, Roche, Sanofi, Siemens, and Singulex. The other authors declare no competing interests.

Peer review

Peer review information.

Nature Communications thanks Rohan Khera, Naeem Ramzan, Robert Miller and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary information, peer review file, reporting summary, rights and permissions.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Cite this article.

Bock, C., Walter, J.E., Rieck, B. et al. Enhancing the diagnosis of functionally relevant coronary artery disease with machine learning. Nat Commun 15 , 5034 (2024). https://doi.org/10.1038/s41467-024-49390-y

Download citation

Received : 05 January 2023

Accepted : 31 May 2024

Published : 12 June 2024

DOI : https://doi.org/10.1038/s41467-024-49390-y

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

By submitting a comment you agree to abide by our Terms and Community Guidelines . If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

Sign up for the Nature Briefing: Translational Research newsletter — top stories in biotechnology, drug discovery and pharma.

An official website of the United States government

Here’s how you know

Official websites use .gov A .gov website belongs to an official government organization in the United States.

Secure .gov websites use HTTPS A lock ( A locked padlock ) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites.

• Heart-Healthy Living
• High Blood Pressure
• Sickle Cell Disease
• Sleep Apnea
• Information & Resources on COVID-19
• The Heart Truth®
• Learn More Breathe Better®
• Blood Diseases and Disorders Education Program
• Publications and Resources
• Blood Disorders and Blood Safety
• Sleep Science and Sleep Disorders
• Lung Diseases
• Health Disparities and Inequities
• Heart and Vascular Diseases
• Precision Medicine Activities
• Obesity, Nutrition, and Physical Activity
• Population and Epidemiology Studies
• Women’s Health
• Research Topics
• Clinical Trials
• All Science A-Z
• Grants and Training Home
• Policies and Guidelines
• Funding Opportunities and Contacts
• Training and Career Development
• Email Alerts
• NHLBI in the Press
• Research Features
• Past Events
• Upcoming Events
• Mission and Strategic Vision
• Divisions, Offices and Centers
• Advisory Committees
• Budget and Legislative Information
• Jobs and Working at the NHLBI
• Contact and FAQs
• NIH Sleep Research Plan
• < Back To Research Topics

Coronary Heart Disease Research

Language switcher.

For almost 75 years, the NHLBI has been at the forefront of improving the nation’s health and reducing the burden of  heart and vascular diseases . Heart disease, including coronary heart disease, remains the leading cause of death in the United States. However, the rate of heart disease deaths has declined by 70% over the past 50 years, thanks in part to NHLBI-funded research. Many current studies funded by the NHLBI focus on discovering genetic associations and finding new ways to prevent and treat the onset of coronary heart disease and associated medical conditions.

NHLBI research that really made a difference

The NHLBI supports a wide range of long-term studies to understand the risk factors of coronary heart disease. These ongoing studies, among others, have led to many discoveries that have increased our understanding of the causes of cardiovascular disease among different populations, helping to shape evidence-based clinical practice guidelines.

• Risk factors that can be changed:  The NHLBI  Framingham Heart Study (FHS)  revealed that cardiovascular disease is caused by modifiable risk factors such as smoking,  high blood pressure ,  obesity , high  cholesterol  levels, and physical inactivity. It is why, in routine physicals, healthcare providers check for high blood pressure, high cholesterol, unhealthy eating patterns, smoking, physical inactivity, and unhealthy weight. The FHS found that cigarette smoking increases the risk of heart disease. Researchers also showed that cardiovascular disease can affect people differently depending on sex or race, underscoring the need to address health disparities. 
• Risk factors for Hispanic/Latino adults:  The  Hispanic Community Health Study/Study of Latinos (HCHS/SOL)  found that heart disease risk factors are widespread among Hispanic/Latino adults in the United States , with 80% of men and 71% of women having at least one risk factor. Researchers also used HCHS/SOL genetic data to explore genes linked with central adiposity (the tendency to have excess body fat around the waist) in Hispanic/Latino adults. Before this study, genes linked with central adiposity, a risk factor for coronary heart disease, had been identified in people of European ancestry. These results showed that those genes also predict central adiposity for Hispanic/Latino communities. Some of the genes identified were more common among people with Mexican or Central/South American ancestry, while others were more common among people of Caribbean ancestry.
• Risk factors for African Americans:  The  Jackson Heart Study (JHS) began in 1997 and includes more than 5,300 African American men and women in Jackson, Mississippi. It has studied genetic and environmental factors that raise the risk of heart problems, especially high blood pressure, coronary heart disease,  heart failure ,  stroke , and  peripheral artery disease (PAD) . Researchers discovered a gene variant in African American individuals that doubles the risk of heart disease. They also found that even small spikes in blood pressure can lead to a higher risk of death. A community engagement component of the JHS is putting 20 years of the study’s findings into action by turning traditional gathering places, such as barbershops and churches, into health information hubs.
• Risk factors for American Indians:  The NHLBI actively supports the  Strong Heart Study , a long-term study that began in 1988 to examine cardiovascular disease and its risk factors among American Indian men and women. The Strong Heart Study is one of the largest epidemiological studies of American Indian people ever undertaken. It involves a partnership with 12 Tribal Nations and has followed more than 8,000 participants, many of whom live in low-income rural areas of Arizona, Oklahoma, and the Dakotas. Cardiovascular disease remains the leading cause of death for American Indian people. Yet the prevalence and severity of cardiovascular disease among American Indian people has been challenging to study because of the small sizes of the communities, as well as the relatively young age, cultural diversity, and wide geographic distribution of the population. In 2019, the NHLBI renewed its commitment to the Strong Heart Study with a new study phase that includes more funding for community-driven pilot projects and a continued emphasis on training and development. Read more about the  goals and key findings  of the Strong Heart Study.

Current research funded by the NHLBI

Within our  Division of Cardiovascular Sciences , the Atherothrombosis and Coronary Artery Disease Branch of its  Adult and Pediatric Cardiac Research Program and the  Center for Translation Research and Implementation Science  oversee much of our funded research on coronary heart disease.

Research funding  

Find  funding opportunities  and  program contacts for research on coronary heart disease. 

Current research on preventing coronary heart disease

• Blood cholesterol and coronary heart disease: The NHLBI supports new research into lowering the risk of coronary heart disease by reducing levels of cholesterol in the blood. High levels of blood cholesterol, especially a type called low-density lipoprotein (LDL) cholesterol, raise the risk of coronary heart disease. However, even with medicine that lowers LDL cholesterol, there is still a risk of coronary heart disease due to other proteins, called triglyceride-rich ApoB-containing lipoproteins (ApoBCLs), that circulate in the blood. Researchers are working to find innovative ways to reduce the levels of ApoBCLs, which may help prevent coronary heart disease and other cardiovascular conditions.
• Pregnancy, preeclampsia, and coronary heart disease risk: NHLBI-supported researchers are investigating the link between developing preeclampsia during pregnancy and an increased risk for heart disease over the lifespan . This project uses “omics” data – such as genomics, proteomics, and other research areas – from three different cohorts of women to define and assess preeclampsia biomarkers associated with cardiovascular health outcomes. Researchers have determined that high blood pressure during pregnancy and low birth weight are predictors of atherosclerotic cardiovascular disease in women . Ultimately, these findings can inform new preventive strategies to lower the risk of coronary heart disease.
• Community-level efforts to lower heart disease risk among African American people: The NHLBI is funding initiatives to partner with churches in order to engage with African American communities and lower disparities in heart health . Studies have found that church-led interventions reduce risk factors for coronary heart disease and other cardiovascular conditions. NHLBI-supported researchers assessed data from more than 17,000 participants across multiple studies and determined that these community-based approaches are effective in lowering heart disease risk factors .

Find more NHLBI-funded studies on  preventing coronary heart disease  on the NIH RePORTER.

Learn about the impact of COVID-19 on your risk of coronary heart disease.

Current research on understanding the causes of coronary heart disease

• Pregnancy and long-term heart disease:  NHLBI researchers are continuing the Nulliparous Pregnancy Outcomes Study: Monitoring Mothers-to-be (nuMoM2b)   study to understand the relationship between pregnancy-related problems, such as gestational hypertension, and heart problems. The study also looks at how problems during pregnancy may increase risk factors for heart disease later in life. NuMoM2b launched in 2010, and long-term studies are ongoing, with the goal of collecting high-quality data and understanding how heart disease develops in women after pregnancy.
• How coronary artery disease affects heart attack risk: NHLBI-funded researchers are investigating why some people with coronary artery disease are more at risk for heart attacks than others. Researchers have found that people with coronary artery disease who have high-risk coronary plaques are more likely to have serious cardiac events, including heart attacks. However, we do not know why some people develop high-risk coronary plaques and others do not. Researchers hope that this study will help providers better identify which people are most at risk of heart attacks before they occur.
• Genetics of coronary heart disease:  The NHLBI supports studies to identify genetic variants associated with coronary heart disease . Researchers are investigating how genes affect important molecular cascades involved in the development of coronary heart disease . This deeper understanding of the underlying causes for plaque buildup and damage to the blood vessels can inform prevention strategies and help healthcare providers develop personalized treatment for people with coronary heart disease caused by specific genetic mutations.

Find more NHLBI-funded studies on understanding the  causes of coronary heart disease  on the NIH RePORTER.

Recent findings suggest that cholesterol-lowering treatment can lower the risk of heart disease complications in people with HIV.

Current research on treatments for coronary heart disease

• Insight into new molecular targets for treatment: NHLBI-supported researchers are investigating the role of high-density lipoprotein (HDL) cholesterol in coronary heart disease and other medical conditions . Understanding how the molecular pathways of cholesterol affect the disease mechanism for atherosclerosis and plaque buildup in the blood vessels of the heart can lead to new therapeutic approaches for the treatment of coronary heart disease. Researchers have found evidence that treatments that boost HDL function can lower systemic inflammation and slow down plaque buildup . This mechanism could be targeted to develop a new treatment approach for coronary heart disease.
• Long-term studies of treatment effectiveness: The NHLBI is supporting the International Study of Comparative Health Effectiveness with Medical and Invasive Approaches (ISCHEMIA) trial EXTENDed Follow-up (EXTEND) , which compares the long-term outcomes of an initial invasive versus conservative strategy for more than 5,000 surviving participants of the original ISCHEMIA trial. Researchers have found no difference in mortality outcomes between invasive and conservative management strategies for patients with chronic coronary heart disease after more than 3 years. They will continue to follow up with participants for up to 10 years. Researchers are also assessing the impact of nonfatal events on long-term heart disease and mortality. A more accurate heart disease risk score will be constructed to help healthcare providers deliver more precise care for their patients.
• Evaluating a new therapy for protecting new mothers: The NHLBI is supporting the Randomized Evaluation of Bromocriptine In Myocardial Recovery Therapy for Peripartum Cardiomyopathy (REBIRTH) , for determining the role of bromocriptine as a treatment for peripartum cardiomyopathy (PPCM). Previous research suggests that prolactin, a hormone that stimulates the production of milk for breastfeeding, may contribute to the development of cardiomyopathy late in pregnancy or the first several months postpartum. Bromocriptine, once commonly used in the United States to stop milk production, has shown promising results in studies conducted in South Africa and Germany. Researchers will enroll approximately 200 women across North America who have been diagnosed with PPCM and assess their heart function after 6 months. 
• Impact of mental health on response to treatment:  NHLBI-supported researchers are investigating how mental health conditions can affect treatment effectiveness for people with coronary heart disease. Studies show that depression is linked to a higher risk for negative outcomes from coronary heart disease. Researchers found that having depression is associated with poor adherence to medical treatment for coronary heart disease . This means that people with depression are less likely to follow through with their heart disease treatment plans, possibly contributing to their chances of experiencing worse outcomes. Researchers are also studying new ways to treat depression in patients with coronary heart disease .

Find more NHLBI-funded studies on  treating coronary heart disease  on the NIH RePORTER.  

Researchers have found no clear difference in patient survival or heart attack risk between managing heart disease through medication and lifestyle changes compared with invasive procedures. 

Coronary heart disease research labs at the NHLBI

• Laboratory of Cardiac Physiology
• Laboratory of Cardiovascular Biology
• Minority Health and Health Disparities Population Laboratory
• Social Determinants of Obesity and Cardiovascular Risk Laboratory
• Laboratory for Cardiovascular Epidemiology and Genomics
• Laboratory for Hemostasis and Platelet Biology

Related coronary heart disease programs

• In 2002, the NHLBI launched  The Heart Truth® ,  the first federally sponsored national health education program designed to raise awareness about heart disease as the leading cause of death in women. The NHLBI and  The Heart Truth®  supported the creation of the Red Dress® as the national symbol for awareness about women and heart disease, and also coordinate  National Wear Red Day ® and  American Heart Month  each February. 
• The  Biologic Specimen and Data Repository Information Coordinating Center (BioLINCC)  facilitates access to and maximizes the scientific value of NHLBI biospecimen and data collections. A main goal is to promote the use of these scientific resources by the broader research community. BioLINCC serves to coordinate searches across data and biospecimen collections and provide an electronic means for requesting additional information and submitting requests for collections. Researchers wanting to submit biospecimen collections to the NHLBI Biorepository to share with qualified investigators may also use the website to initiate the application process. 
• Our  Trans-Omics for Precision Medicine (TOPMed) Program  studies the ways genetic information, along with information about health status, lifestyle, and the environment, can be used to predict the best ways to prevent and treat heart, lung, blood, and sleep disorders. TOPMed specifically supports NHLBI’s  Precision Medicine Activities. 
• NHLBI  population and epidemiology studies  in different groups of people, including the  Atherosclerosis Risk in Communities (ARIC) Study , the  Multi-Ethnic Study of Atherosclerosis (MESA) , and the  Cardiovascular Health Study (CHS) , have made major contributions to understanding the causes and prevention of heart and vascular diseases, including coronary heart disease.
• The  Cardiothoracic Surgical Trials Network (CTSN)  is an international clinical research enterprise that studies  heart valve disease ,  arrhythmias , heart failure, coronary heart disease, and surgical complications. The trials span all phases of development, from early translation to completion, and have more than 14,000 participants. The trials include six completed randomized clinical trials, three large observational studies, and many other smaller studies.

Learn how heart disease may be different for women than for men.

Explore more NHLBI research on coronary heart disease

The sections above provide you with the highlights of NHLBI-supported research on coronary heart disease. You can explore the full list of NHLBI-funded studies on the NIH RePORTER .

To find more studies:

• Type your search words into the  Quick Search  box and press enter. 
• Check  Active Projects  if you want current research.
• Select the  Agencies  arrow, then the  NIH  arrow, then check  NHLBI .

If you want to sort the projects by budget size — from the biggest to the smallest — click on the  FY Total Cost by IC  column heading.

• Case report
• Open access
• Published: 16 June 2024

Spontaneous intercostal artery bleeding occurring simultaneously in numerous vessels during antithrombotic therapy with mechanical circulatory support: a case report

• Kazuto Ohtaka 1 , 2 ,
• Setsuyuki Ohtake 1 ,
• Yu Ishii 1 ,
• Saya Kaku 1 ,
• Yuta Takeuchi 1 ,
• Tomoko Mizota 1 ,
• Yoshiyuki Yamamura 1 ,
• Masaomi Ichinokawa 1 ,
• Tatsuya Yoshioka 1 ,
• Eiji Tamoto 1 ,
• Katsuhiko Murakawa 1 ,
• Koichi Ono 1 &
• Tatsuya Kato 2  

Journal of Medical Case Reports volume  18 , Article number:  280 ( 2024 ) Cite this article

137 Accesses

1 Altmetric

Metrics details

Intercostal artery bleeding often occurs in a single vessel; in rare cases, it can occur in numerous vessels, making it more difficult to manage.

Case presentation

A 63-year-old Japanese man was admitted to the emergency department owing to sudden chest and back pain, dizziness, and nausea. Emergency coronary angiography revealed myocardial infarction secondary to right coronary artery occlusion. After intra-aortic balloon pumping, percutaneous coronary intervention was performed in the right coronary artery. At 12 hours following percutaneous coronary intervention, the patient developed new-onset left anterior chest pain and hypotension. Contrast-enhanced computed tomography revealed 15 sites of contrast extravasation within a massive left extrapleural hematoma. Emergency angiography revealed contrast leakage in the left 6th to 11th intercostal arteries; hence, transcatheter arterial embolization was performed. At 2 days after transcatheter arterial embolization, his blood pressure subsequently decreased, and contrast-enhanced computed tomography revealed the re-enlargement of extrapleural hematoma with multiple sites of contrast extravasation. Emergency surgery was performed owing to persistent bleeding. No active arterial hemorrhage was observed intraoperatively. Bleeding was observed in various areas of the chest wall, and an oxidized cellulose membrane was applied following ablation and hemostasis. The postoperative course was uneventful.

We report a case of spontaneous intercostal artery bleeding occurring simultaneously in numerous vessels during antithrombotic therapy with mechanical circulatory support that was difficult to manage. As bleeding from numerous vessels may occur during antithrombotic therapy, even without trauma, appropriate treatments, such as transcatheter arterial embolization and surgery, should be selected in patients with such cases.

Peer Review reports

Intercostal artery bleeding arises from vessel fragility induced by various underlying conditions, including neurofibromatosis type 1, coarctation of the aorta, systemic lupus erythematosus, alcoholic cirrhosis, and trauma leading to aneurysm formation and rupture [ 1 , 2 , 3 , 4 , 5 , 6 ]. Intercostal artery bleeding can cause massive hemothorax, chest wall hematomas, abdominal wall hematomas, and paravertebral hematomas, some of which can be fatal [ 7 , 8 , 9 ]. It is often diagnosed using contrast-enhanced computed tomography (CT), which shows contrast extravasation [ 1 ]. While single-vessel bleeding is common, simultaneous bleeding from numerous vessels is rare [ 1 , 10 ]. We report a case of a patient experiencing spontaneous intercostal artery bleeding from numerous vessels during antithrombotic therapy, making it difficult to manage.

A 63-year-old Japanese man with type 2 diabetes mellitus who self-discontinued treatment and had a 20-pack-year smoking history was admitted to the emergency room owing to sudden chest and back pain, dizziness, and nausea. Although initial chest radiography and CT scans showed no anomalies, an electrocardiogram showed ST-segment elevation in leads II, III, and aV F . Subsequent coronary angiography (CAG) exposed three-vessel coronary artery disease, specifically myocardial infarction owing to right coronary artery occlusion. To address hemodynamic instability, such as systolic blood pressure failing to 50 mmHg, intra-aortic balloon pumping (IABP) and a temporary pacemaker were employed in addition to vasopressor administration. Percutaneous coronary intervention (PCI) was performed only on the right coronary artery, which was considered to be the culprit lesion. The patient was then transferred to the intensive care unit (ICU). Dual antiplatelet therapy (DAPT) with 100 mg of aspirin and 3.75 mg of prasugrel hydrochloride per day was initiated, and heparin was administered at a rate of 15,000 units per day for IABP, with checking the coagulation function every 8 hours.

At 12 hours following PCI, the patient encountered left anterior chest pain, leading to reduced systolic blood pressure to 60 mmHg and hemoglobin levels to 70 g/L. Intubation was performed for pain relief after fluid and blood transfusions raised his blood pressure. The patient had a prothrombin activity rate of 89% and an activated partial thromboplastin time of 90.5 seconds. Contrast-enhanced CT scans after hemodynamic stabilization revealed a sizable extrapleural hematoma with 15 sites of contrast extravasation, that were suspected to be numerous intercostal bleeding (Fig.  1 ). While transcatheter arterial embolization (TAE) was planned, it was initially challenging owing to the position of the IABP balloon and the target intercostal artery. After IABP removal, veno-arterial extracorporeal membrane oxygenation (VA-ECMO) was introduced. A contrast leakage was noted in the left 6th to 11th intercostal arteries, leading to successful embolization (Fig.  2 ). Following TAE, an IABP was reinserted, and heparin was administered, maintaining clotting time. Hemodynamic stability was restored after blood transfusion.

Contrast-enhanced computed tomography scan showing a large left extrapleural hematoma and multiple-vessel bleeding within the hematoma. Yellow numbers indicate the rib numbers. White arrows indicate the parietal pleura. White arrowheads indicate contrast extravasation

Emergency angiography showing contrast leakage from the left 6th to 11th intercostal arteries. Transcatheter arterial embolization was performed on these vessels. Black arrows indicate contrast extravasation

However, 2 days later, blood pressure dropped, and anemia worsened. The patient had a prothrombin activity rate of 83% and an activated partial thromboplastin time of 48.4 seconds. Then, CT scans showed hematoma re-enlargement, multiple contrast extravasation sites, and lung atelectasis (Fig.  3 ). An emergency thoracotomy was performed owing to persistent bleeding. The large hematoma was excised, with no active arterial hemorrhage observed. Various chest wall areas exhibited bleeding, managed through ablation and hemostasis. The procedure lasted 81 min, with 2616 mL blood loss. VA-ECMO ceased on postoperative day (POD) 1. POD2 saw IABP removal, POD3 extubation, and POD4 chest drain removal. On POD37, after extended rehabilitation, the patient was discharged. A month later, coronary artery bypass grafting addressed the remaining lesions, including left main coronary artery lesion.

Contrast-enhanced computed tomography scan following transcatheter arterial embolization, revealing a re-enlarged extrapleural hematoma, multiple-vessel bleeding (white arrowheads), and atelectasis of the left lung (white arrows). Yellow numbers indicate the rib numbers. White arrows indicate the left lung. White arrowheads indicate contrast extravasation

Intercostal artery bleeding can also be precipitated by medical procedures, such as dialysis, transcatheter aortic valve implantation, ultrasound-guided liver biopsy, PCI, and IABP [ 6 , 11 , 12 ]. In this case, the patient initially presented with three-vessel coronary artery disease identified through CAG. Subsequently, a pacemaker and IABP were implanted to manage hemodynamic instability, followed by PCI. While bleeding complications from arterial guidewire manipulation during PCI have been reported in 0.2–0.5% of patients, these incidents typically occur during the procedure [ 13 ]. In contrast, our patient experienced shock approximately 12 hours post-PCI, suggesting a different sequence of events. A similar case reported by Shiraishi et al . described left hemothorax and shock occurring 8 hours following IABP placement [ 14 ]. In their study, a catheter twist in the descending aorta led to perforation, resulting in bleeding from one site originating from the aorta. This contrasts with the present case, where no procedural complications were observed, classifying the bleeding as spontaneous. In this case report, the patient received post-PCI DAPT and was managed with heparin for IABP and VA-ECMO support, both during initial bleeding and after TAE. Anticoagulation has been linked to inducing spontaneous hemothorax [ 15 , 16 ]. DAPT carries a higher risk of life-threatening bleeding events compared with single antiplatelet agents, and the risk escalates with heparin use [ 17 ]. Additionally, soft tissue bleeding is more prone to occur during mechanical circulatory support [ 18 ]. These facts suggest that antithrombotic therapy significantly contributed to the bleeding in this patient. Anticoagulation-related bleeding during mechanical circulatory support often proves fatal, underscoring the necessity of discontinuing antithrombotic therapy and improving coagulation function.

Previous case reports of intercostal artery bleeding, except those stemming from trauma, have predominantly featured bleeding from a single vessel, with instances of dual- or triple-vessel bleeding being exceedingly rare [ 1 , 10 ]. In this patient, contrast-enhanced CT showed 15 extravasation sites suggesting bleeding in numerous vessels, prompting TAE to embolize the six intercostal arteries, which is the most commonly reported procedure to date. As the patient was presumed to have experienced continued bleeding, open thoracotomy hemostasis was required. It remains uncertain whether all bleeding originated solely from the intercostal artery of the present patient.

The primary treatment for intercostal artery bleeding is TAE, which has demonstrated relatively favorable outcomes [ 4 , 6 , 10 ]. On the contrary, emergency open thoracotomy has shown less favorable results [ 19 , 20 ]. Emergency open thoracotomy presents challenges in identifying the precise source of bleeding during the acute phase, often necessitating subsequent reoperation. Therefore, TAE is generally favored over emergency open thoracotomy. Tanaka et al . proposed the surgical removal of a hematoma after stabilization with TAE [ 21 ]. Without uncontrolled or massive bleeding, conservative management may also be a viable option [ 22 ]. In the present patient, the bleeding occurred from multiple vessels, posing challenges in its management. In such cases, it may be necessary to use a combination of several treatment approaches.

We report a case of spontaneous intercostal artery bleeding occurring simultaneously in numerous vessels during antithrombotic therapy with mechanical circulatory support, possibly representing the most extensive occurrence of such bleeding reported thus far. Because bleeding can occur from numerous vessels even without trauma, the findings of contrast-enhanced CT should be carefully interpreted to identify bleeding sites and determine the optimal treatment strategy. Surgical intervention may be considered if CT findings show extensive atelectasis with a massive hematoma or multiple-vessel bleeding possibly owing to nonarterial bleeding.

Availability of data and materials

All data generated or analyzed during this study are included in this published article.

Abbreviations

Computed tomography

Coronary angiography

Intra-aortic balloon pumping

Percutaneous coronary intervention

Intensive care unit

Dual antiplatelet therapy

• Transcatheter arterial embolization

Venoarterial extracorporeal membrane oxygenation

Postoperative day

Wightman SC, Wang Y, Rohr AM, Greene CL, Hwang GL, Watkins AC, et al . Spontaneous bleeding from multiple intercostal arteries in a patient with coarctation of the aorta. Ann Thorac Surg. 2020;110:e95–7. https://doi.org/10.1016/j.athoracsur.2019.12.042 .

Article   PubMed   Google Scholar  

Fukuda W, Taniguchi S, Fukuda I. Endovascular treatment of ruptured intercostal arteriovenous fistulas associated with neurofibromatosis type 1. Ann Vasc Dis. 2012;5:109–12. https://doi.org/10.3400/avd.cr.11.00078 .

Article   PubMed   PubMed Central   Google Scholar  

Lu CC, Chen CH, Yeh SF, Lai JH, Chang DM. A spontaneous intercostal artery hemorrhage in systemic lupus erythematosus. Rheumatol Int. 2012;32:829–31. https://doi.org/10.1007/s00296-011-1826-x .

Article   CAS   PubMed   Google Scholar  

Rahi MS, Pednekar P, Parmar G, Keibel L, Gunasekaran K, Amoah K, et al . Spontaneous intercostal artery bleeding in a patient with alcohol-induced liver cirrhosis. Clin Case Rep. 2021;9: e04613. https://doi.org/10.1002/ccr3.4613 .

Gutierrez Romero DF, Barrufet M, Lopez-Rueda A, Burrel M. Ruptured intercostal artery pseudoaneurysm in a patient with blunt thoracic: trauma diagnosis and management. BMJ Case Rep. 2014. https://doi.org/10.1136/bcr-2013-202019 .

Durey A, Kim YS, Kim AJ. Spontaneous intercostal artery bleeding in a hemodialysis patient. Hemodial Int. 2017;21:e76–8. https://doi.org/10.1111/hdi.12576 .

Liu C, Ran R, Li X, Liu G, Wang C, Li J. Massive hemothorax caused by intercostal artery pseudoaneurysm: a case report. J Cardiothorac Surg. 2021;16:156. https://doi.org/10.1186/s13019-021-01548-1 .

Mathew R, Abdullah S, Renfrew I. Massive abdominal wall haematoma and haemothorax due to spontaneous rupture of an intercostal artery. Emerg Med J. 2008;25:608. https://doi.org/10.1136/emj.2007.057497 .

Izumoto S, Abe T, Koroki T, Furukoji E, Masuda R, Ochiai H. A 48-year-old man presenting as an emergency with severe back pain a large anterior paravertebral hematoma and spontaneous rupture of the right 9th intercostal artery successfully managed by transcatheter arterial embolization: a case report. Am J Case Rep. 2022;23: e934173. https://doi.org/10.12659/AJCR.934173 .

Moon JM, Lee SC, Chun BJ. Spontaneous intercostal artery bleeding. Emerg Med J. 2008;25:53–4. https://doi.org/10.1136/emj.2007.052548 .

Lenders G, Van Schil P, Rodrigus I, Bosmans J. Intercostal artery pseudoaneurysm: a rare complication of transaortic transcatheter aortic valve implantation. Interact Cardiovasc Thorac Surg. 2012;15:550–2. https://doi.org/10.1093/icvts/ivs188 .

Vajtai Z, Roy N. Intercostal artery pseudoaneurysm after ultrasound-guided liver biopsy: a case report and review of the literature. Ultrasound Q. 2015;31:63–5. https://doi.org/10.1097/RUQ.0000000000000074 .

Hess CN, Rao SV, McCoy LA, Neely ML, Singh M, Spertus JA, et al . Identification of hospital outliers in bleeding complications after percutaneous coronary intervention. Circ Cardiovasc Qual Outcomes. 2015;8:15–22. https://doi.org/10.1161/CIRCOUTCOMES.113.000749 .

Shiraishi R, Okazaki Y, Naito K, Itoh T. Perforation of the descending aorta by the tip of an intra-aortic balloon pump catheter. Circ J. 2002;66:423–4. https://doi.org/10.1253/circj.66.423 .

Zeiler J, Idell S, Norwood S, Cook A. Hemothorax: a review of the literature. Clin Pulm Med. 2020;27:1–12.

Patrini D, Panagiotopoulos N, Pararajasingham J, Gvinianidze L, Iqbal Y, Lawrence DR. Etiology and management of spontaneous haemothorax. J Thorac Dis. 2015;7:520–6.

PubMed   PubMed Central   Google Scholar  

Toyoda K, Yasaka M, Iwade K, Nagata K, Koretsune Y, Sakamoto T, Bleeding with Antithrombotic Therapy (BAT) Study Group, et al . Dual antithrombotic therapy increases severe bleeding events in patients with stroke and cardiovascular disease: a prospective, multicenter, observational study. Stroke. 2008;39:1740–5.

Taniguchi H, Ikeda T, Takeuchi I, Ichiba S. Iliopsoas hematoma in patients undergoing venovenous ECMO. Am J Circ Care. 2021;30:55–63.

Article   Google Scholar  

Dua A, Dua A, Jechow S, Desai SS, Kuy SR. Idiopathic spontaneous rupture of an intercostal artery. WMJ. 2014;113:116–8.

PubMed   Google Scholar  

Aizawa K, Iwashita C, Saito T, Misawa Y. Spontaneous rupture of an intercostal artery in a patient with neurofibromatosis type 1. Interact Cardiovasc Thorac Surg. 2010;10:128–30. https://doi.org/10.1510/icvts.2009.222125 .

Tanaka Y, Haratake N, Kinoshita F, Takenaka T, Tagawa T, Mori M. Spontaneous hemopneumothorax with a ruptured aneurysm in the second intercostal artery: report of a case. Gen Thorac Cardiovasc Surg. 2021;69:1133–6. https://doi.org/10.1007/s11748-021-01620-6 .

Ishida A, Hiraoka A, Chikazawa G, Maeda K, Yoshitaka H. Spontaneous intercostal arterial rupture restrained by conservative management. Ann Vasc Dis. 2014;7:430–2. https://doi.org/10.3400/avd.cr.14-00087 .

Download references

Acknowledgements

We would like to thank Dr. Shohei Hakozaki, Department of Cardiology, Obihiro Kosei General Hospital, for assisting with patient management.

The authors did not receive any funds related to this case report.

Author information

Authors and affiliations.

Department of Surgery, Obihiro Kosei General Hospital, West 14 South 10, Obihiro, Hokkaido, 080-0024, Japan

Kazuto Ohtaka, Setsuyuki Ohtake, Yu Ishii, Saya Kaku, Yuta Takeuchi, Tomoko Mizota, Yoshiyuki Yamamura, Masaomi Ichinokawa, Tatsuya Yoshioka, Eiji Tamoto, Katsuhiko Murakawa & Koichi Ono

Department of Thoracic Surgery, Hokkaido University Hospital, West 5 North 14, Kita-Ku, Sapporo, Hokkaido, 060-8648, Japan

Kazuto Ohtaka & Tatsuya Kato

You can also search for this author in PubMed   Google Scholar

Contributions

KO was the main surgeon and drafted the manuscript; SO assisted with the surgery and helped draft the manuscript; YI, SK, YT, TM, YY, MI, TY, ET, KM, and KO assisted with patient care and drafted the manuscript; TK reviewed the manuscript. All authors have read and approved the final manuscript.

Corresponding author

Correspondence to Kazuto Ohtaka .

Ethics declarations

Ethics approval and consent to participate.

Not applicable.

Consent for publication

Written informed consent was obtained from the patient for publication of this case report and any accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal.

Competing interests

The authors declare that they have no competing interests.

Additional information

Publisher’s note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ . The Creative Commons Public Domain Dedication waiver ( http://creativecommons.org/publicdomain/zero/1.0/ ) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Cite this article.

Ohtaka, K., Ohtake, S., Ishii, Y. et al. Spontaneous intercostal artery bleeding occurring simultaneously in numerous vessels during antithrombotic therapy with mechanical circulatory support: a case report. J Med Case Reports 18 , 280 (2024). https://doi.org/10.1186/s13256-024-04602-3

Download citation

Received : 28 January 2024

Accepted : 29 May 2024

Published : 16 June 2024

DOI : https://doi.org/10.1186/s13256-024-04602-3

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Spontaneous intercostal artery bleeding
• Extravasation
• Extrapleural hematoma

Journal of Medical Case Reports

ISSN: 1752-1947

• Submission enquiries: Access here and click Contact Us
• General enquiries: [email protected]

• - Google Chrome

Intended for healthcare professionals

• My email alerts
• BMA member login
• Username * Password * Forgot your log in details? Need to activate BMA Member Log In Log in via OpenAthens Log in via your institution

Search form

• Advanced search
• Search responses
• Search blogs
• News & Views
• Dual antiplatelet...

Dual antiplatelet therapy after coronary artery bypass surgery

Linked research.

Antiplatelet therapy after coronary artery bypass surgery

• Related content
• Peer review
• Sigrid Sandner , associate professor of cardiac surgery 1 2
• 1 Department of Cardiac Surgery, Medical University of Vienna, Austria
• 2 Department of Cardiothoracic Surgery, Weill Cornell Medicine, New York, NY, USA
• sigrid.sandner{at}meduniwien.ac.at

Dual therapy is linked to lower risk of major cardiovascular events over five years

Antiplatelet therapy is an integral part of secondary prevention for patients with coronary artery disease, as inhibition of platelet aggregation reduces the risk of coronary plaque thrombosis and subsequent cardiovascular events. 1 Among patients with coronary artery disease undergoing coronary artery bypass graft surgery, antiplatelet therapy also prevents early saphenous vein graft occlusion. 2 The saphenous vein graft is the most common coronary artery bypass graft and is used in more than 90% of procedures in the US owing to its general availability and easy deployment. 3 However, graft patency remains suboptimal, with a high risk of occlusion early after surgery which decreases after the first year. 2 During surgical manipulation, the graft endothelium incurs mechanical injury and transient ischemia, triggering a platelet driven thrombotic cascade that is responsible for the high early occlusion rates (up to 20% within the first year after surgery). 4

Studies conducted several decades ago showed that aspirin at a dose of 100-325 mg once daily significantly decreased the …

Log in using your username and password

BMA Member Log In

If you have a subscription to The BMJ, log in:

• Need to activate
• Log in via institution
• Log in via OpenAthens

Log in through your institution

Subscribe from £184 *.

Subscribe and get access to all BMJ articles, and much more.

* For online subscription

Access this article for 1 day for: £33 / $40 / €36 ( excludes VAT )

You can download a PDF version for your personal record.

Buy this article

• Patient Care & Health Information
• Diseases & Conditions
• Coronary artery disease

Coronary artery disease (CAD) is a common type of heart disease. It affects the main blood vessels that supply blood to the heart, called the coronary arteries. In CAD, there is reduced blood flow to the heart muscle. A buildup of fats, cholesterol and other substances in and on the artery walls, a condition called atherosclerosis, usually causes coronary artery disease. The buildup, called plaque, makes the arteries narrow.

Coronary artery disease often develops over many years. Symptoms are from the lack of blood flow to the heart. They may include chest pain and shortness of breath. A complete blockage of blood flow can cause a heart attack.

Treatment for coronary artery disease may include medicines and surgery. Eating a nutritious diet, getting regular exercise and not smoking can help prevent coronary artery disease and the conditions that can cause it.

Coronary artery disease also may be called coronary heart disease.

• What is coronary artery disease? A Mayo Clinic cardiologist explains.

Stephen Kopecky, M.D., talks about the risk factors, symptoms and treatment of coronary artery disease (CAD). Learn how lifestyle changes can lower your risk.

{Music playing}

Stephen Kopecky, M.D., Cardiovascular Disease, Mayo Clinic: I'm Dr. Stephen Kopecky, a cardiologist at Mayo Clinic. In this video, we'll cover the basics of coronary artery disease. What is it? Who gets it? The symptoms, diagnosis and treatment. Whether you're looking for answers for yourself or someone you love, we're here to give you the best information available.

Coronary artery disease, also called CAD, is a condition that affects your heart. It is the most common heart disease in the United States. CAD happens when coronary arteries struggle to supply the heart with enough blood, oxygen and nutrients. Cholesterol deposits, or plaques, are almost always to blame. These buildups narrow your arteries, decreasing blood flow to your heart. This can cause chest pain, shortness of breath or even a heart attack. CAD typically takes a long time to develop. So often, patients don't know that they have it until there's a problem. But there are ways to prevent coronary artery disease, and ways to know if you're at risk and ways to treat it.

Who gets it?

Anyone can develop CAD . It begins when fats, cholesterols and other substances gather along the walls of your arteries. This process is called atherosclerosis. It's typically no cause for concern. However, too much buildup can lead to a blockage, obstructing blood flow. There are a number of risk factors, common red flags, that can contribute to this and ultimately lead to coronary artery disease. First, getting older can mean more damaged and narrowed arteries. Second, men are generally at a greater risk. But the risk for women increases after menopause. Existing health conditions matter, too. High blood pressure can thicken your arteries, narrowing your blood flow. High cholesterol levels can increase the rate of plaque buildup. Diabetes is also associated with higher risk, as is being overweight. Your lifestyle plays a large role as well. Physical inactivity, long periods of unrelieved stress in your life, an unhealthy diet and smoking can all increase your risk. And finally, family history. If a close relative was diagnosed at an early age with heart disease, you're at a greater risk. All these factors together can paint a picture of your risk for developing CAD .

What are the symptoms?

When coronary arteries become narrow, the heart doesn't get enough oxygen-rich blood. Remember, unlike most pumps, the heart has to pump its own energy supply. It's working harder with less. And you may begin to notice these signs and symptoms of pressure or tightness in your chest. This pain is called angina. It may feel like somebody is standing on your chest. When your heart can't pump enough blood to meet your body's needs, you might develop shortness of breath or extreme fatigue during activities. And if an artery becomes totally blocked, it leads to a heart attack. Classic signs and symptoms of a heart attack include crushing, substernal chest pain, pain in your shoulders or arms, shortness of breath, and sweating. However, many heart attacks have minimal or no symptoms and are found later during routine testing.

How is it diagnosed?

Diagnosing CAD starts by talking to your doctor. They'll be able to look at your medical history, do a physical exam and order routine blood work. Depending on that, they may suggest one or more of the following tests: an electrocardiogram or ECG, an echocardiogram or soundwave test of the heart, stress test, cardiac catheterization and angiogram, or a cardiac CT scan.

How is it treated?

Treating coronary artery disease usually means making changes to your lifestyle. This might be eating healthier foods, exercising regularly, losing excess weight, reducing stress or quitting smoking. The good news is these changes can do a lot to improve your outlook. Living a healthier life translates to having healthier arteries. When necessary, treatment could involve drugs like aspirin, cholesterol-modifying medications, beta-blockers, or certain medical procedures like angioplasty or coronary artery bypass surgery.

Discovering you have coronary artery disease can be overwhelming. But be encouraged. There are things you can do to manage and live with this condition. Reducing cholesterol, lowering blood pressure, quitting tobacco, eating healthier, exercising and managing your stress can make a world of difference. Better heart health starts by educating yourself. So don't be afraid to seek out information and ask your doctors about coronary artery disease. If you'd like to learn even more about this condition, watch our other related videos or visit Mayoclinic.org. We wish you well.

Symptoms of coronary artery disease happen when the heart doesn't get enough oxygen-rich blood. Coronary artery disease symptoms may include:

• Chest pain, called angina. You may feel squeezing, pressure, heaviness, tightness or pain in the chest. It may feel like somebody is standing on your chest. The chest pain usually affects the middle or left side of the chest. Activity or strong emotions can trigger angina. There are different types of angina. The type depends on the cause and whether rest or medicine makes symptoms better. In some people, especially women, the pain may be brief or sharp and felt in the neck, arm or back.
• Shortness of breath. You may feel like you can't catch your breath.
• Fatigue . If the heart can't pump enough blood to meet your body's needs, you may feel unusually tired.

Symptoms of coronary artery disease may not be noticed at first. Sometimes symptoms only happen when the heart is beating hard, such as during exercise. As the coronary arteries continue to narrow, symptoms can get more severe or frequent.

A completely blocked coronary artery will cause a heart attack. Common heart attack symptoms include:

• Chest pain that may feel like pressure, tightness, squeezing or aching.
• Pain or discomfort that spreads to the shoulder, arm, back, neck, jaw, teeth or sometimes the upper belly.
• Cold sweats.
• Shortness of breath.
• Lightheadedness or sudden dizziness.

Chest pain is usually the most common symptom of heart attack. But for some people, such as women, the elderly and those with diabetes, symptoms may seem unrelated to a heart attack. For example, they may have nausea or a very brief pain in the neck or back. Some people having a heart attack don't notice symptoms.

When to see a doctor

If you think you're having a heart attack, immediately call 911 or your local emergency number. If you don't have access to emergency medical services, have someone drive you to the nearest hospital. Drive yourself only as a last option.

Smoking or having high blood pressure, high cholesterol, diabetes, obesity or a strong family history of heart disease makes you more likely to get coronary artery disease. If you're at high risk of coronary artery disease, talk to your healthcare professional. You may need tests to check for narrowed arteries and coronary artery disease.

There is a problem with information submitted for this request. Review/update the information highlighted below and resubmit the form.

From Mayo Clinic to your inbox

Sign up for free and stay up to date on research advancements, health tips, current health topics, and expertise on managing health. Click here for an email preview.

Error Email field is required

Error Include a valid email address

To provide you with the most relevant and helpful information, and understand which information is beneficial, we may combine your email and website usage information with other information we have about you. If you are a Mayo Clinic patient, this could include protected health information. If we combine this information with your protected health information, we will treat all of that information as protected health information and will only use or disclose that information as set forth in our notice of privacy practices. You may opt-out of email communications at any time by clicking on the unsubscribe link in the e-mail.

Thank you for subscribing!

You'll soon start receiving the latest Mayo Clinic health information you requested in your inbox.

Sorry something went wrong with your subscription

Please, try again in a couple of minutes

Development of atherosclerosis

If there's too much cholesterol in the blood, the cholesterol and other substances may form deposits called plaque. Plaque can cause an artery to become narrowed or blocked. If a plaque ruptures, a blood clot can form. Plaque and blood clots can reduce blood flow through an artery.

Coronary artery disease is caused by the buildup of fats, cholesterol and other substances in and on the walls of the heart arteries. This condition is called atherosclerosis. The buildup is called plaque. Plaque can cause the arteries to narrow, blocking blood flow. The plaque also can burst, causing a blood clot.

Some causes of atherosclerosis and coronary artery disease are:

• Diabetes or insulin resistance.
• High blood pressure.
• Lack of exercise.
• Smoking or tobacco use.

Risk factors

Coronary artery disease is common.

Coronary artery disease risk factors you can't control include:

• Age. Getting older increases the risk of damaged and narrowed arteries.
• Birth sex. Men are generally at greater risk of coronary artery disease. However, the risk for women increases after menopause.
• Family history. A family history of heart disease makes you more likely to get coronary artery disease. This is especially true if a parent, brother, sister or child got heart disease at an early age. The risk is highest if your father or a brother had heart disease before age 55 or if your mother or a sister developed it before age 65.

Coronary artery disease risk factors you can control are:

• Smoking. If you smoke, quit. Smoking is bad for heart health. People who smoke have a much greater risk of heart disease. Breathing in secondhand smoke also increases the risk.
• High blood pressure. Uncontrolled high blood pressure can make arteries hard and stiff. This can lead to atherosclerosis, which causes coronary artery disease.
• Cholesterol. Too much "bad" cholesterol in the blood can increase the risk of atherosclerosis. "Bad" cholesterol is called low-density lipoprotein (LDL) cholesterol. Not enough "good" cholesterol, called high-density lipoprotein (HDL) cholesterol, also leads to atherosclerosis.
• Diabetes. Diabetes increases the risk of coronary artery disease. Type 2 diabetes and coronary artery disease share some risk factors, such as obesity and high blood pressure.
• Obesity. Too much body fat is bad for overall health. Obesity can lead to type 2 diabetes and high blood pressure. Ask your healthcare team what a healthy weight is for you.
• Chronic kidney disease. Having long-term kidney disease increases the risk of coronary artery disease.
• Not getting enough exercise. Physical activity is important for good health. A lack of exercise is linked to coronary artery disease and some of its risk factors.
• A lot of stress . Emotional stress may damage the arteries and worsen other risk factors for coronary artery disease.
• Unhealthy diet. Eating foods with a lot of saturated fat, trans fat, salt and sugar can increase the risk of coronary artery disease.
• Alcohol use. Heavy alcohol use can lead to heart muscle damage. It also can worsen other risk factors of coronary artery disease.
• Amount of sleep. Too little sleep and too much sleep both have been linked to an increased risk of heart disease.

Risk factors often happen together. One risk factor may trigger another. When grouped together, some risk factors make you even more likely to develop coronary artery disease. For example, metabolic syndrome is a group of conditions that includes high blood pressure, high blood sugar, too much body fat around the waist and high triglyceride levels. Metabolic syndrome increases the risk of coronary artery disease.

Other possible risk factors for coronary artery disease may include:

• Breathing pauses during sleep, called obstructive sleep apnea. This condition causes breathing to stop and start during sleep. It can cause sudden drops in blood oxygen levels. The heart must work harder to pump blood. Blood pressure goes up.
• Increased high-sensitivity C-reactive protein (hs-CRP). This protein appears in higher than usual amounts when there's inflammation somewhere in the body. High hs-CRP levels may be a risk factor for heart disease. It's thought that as coronary arteries narrow, the level of hs-CRP in the blood goes up.
• High triglycerides. This is a type of fat in the blood. High levels may raise the risk of coronary artery disease, especially for women.
• High levels of homocysteine. Homocysteine is a substance that the body uses to make protein and to build and maintain tissue. But high levels of homocysteine may raise the risk of coronary artery disease.
• Preeclampsia. This pregnancy complication causes high blood pressure and increased protein in the urine. It can lead to a higher risk of heart disease later in life.
• Other pregnancy complications. Diabetes or high blood pressure during pregnancy are known risk factors for coronary artery disease.
• Certain autoimmune diseases. People who have conditions such as rheumatoid arthritis and lupus have an increased risk of atherosclerosis.

Complications

Complications of coronary artery disease may include:

• Chest pain, also called angina. This is a symptom of coronary artery disease. But it also can be a complication of worsening CAD. The chest pain happens when arteries narrow and the heart doesn't get enough blood.
• Heart attack. A heart attack can happen if atherosclerosis causes a blood clot. A clot can block blood flow. The lack of blood can damage the heart muscle. The amount of damage depends in part on how quickly you are treated.
• Heart failure. Narrowed arteries in the heart or high blood pressure can slowly make the heart weak or stiff. This can make it harder for the heart to pump blood.
• Irregular heart rhythms, called arrhythmias. If the heart doesn't get enough blood, changes in heart signaling can happen. This can cause irregular heartbeats.

Coronary artery disease care at Mayo Clinic

Living with coronary artery disease?

Connect with others like you for support and answers to your questions in the Heart & Blood Health support group on Mayo Clinic Connect, a patient community.

Heart & Blood Health Discussions

106 Replies Wed, Jun 19, 2024

341 Replies Wed, Jun 19, 2024

385 Replies Tue, Jun 04, 2024

• Ferri FF. Coronary artery disease. In: Ferri's Clinical Advisor 2022. Elsevier; 2022. https://www.clinicalkey.com. Accessed March 8, 2022.
• Coronary heart disease. National Heart, Lung, and Blood Institute. https://www.nhlbi.nih.gov/health-topics/coronary-heart-disease. March 8, 2022.
• Usatine RP, et al., eds. Coronary artery disease. In: Color Atlas and Synopsis of Heart Failure. McGraw Hill; 2019.
• Wilson PWF. Overview of the possible risk factors for cardiovascular disease. https://www.uptodate.com/contents/search. Accessed March 8, 2022.
• Masjedi MS, et al. Effects of flaxseed on blood lipids in healthy and dyslipidemic subjects: A systematic review and meta-analysis of randomized controlled trials. Current Problems in Cardiology. 2021; doi:10.1016/j.cpcardiol.2021.100931.
• Riaz H, et al. Association between obesity and cardiovascular outcomes: A systematic review and meta-analysis of mendelian randomization studies. JAMA Network Open. 2018; doi:10.1001/jamanetworkopen.2018.3788.
• Physical Activity Guidelines for Americans. 2nd ed. U.S. Department of Health and Human Services. https://health.gov/our-work/physical-activity/current-guidelines. Accessed March 8, 2022.
• Your guide to lowering your cholesterol with therapeutic lifestyle changes (TLC). National Heart, Lung, and Blood Institute. https://www.nhlbi.nih.gov/health-topics/all-publications-and-resources/your-guide-lowering-cholesterol-therapeutic-lifestyle. Accessed March 24, 2022.
• Rethinking drinking. National Institute on Alcohol Abuse and Alcoholism. https://www.rethinkingdrinking.niaaa.nih.gov/. Accessed March 24, 2022.
• 2015-2020 Dietary Guidelines for Americans. U.S. Department of Health and Human Services and U.S. Department of Agriculture. https://health.gov/our-work/food-nutrition/2015-2020-dietary-guidelines/guidelines. Accessed March 24, 2022.
• Omega-3 supplements: In depth. National Center for Complementary and Integrative Health. https://www.nccih.nih.gov/health/omega3-supplements-in-depth. Accessed March 8, 2022.
• Lopez-Jimenez F (expert opinion). Mayo Clinic. May 9, 2024.
• Siscovick DS, et al. Omega-3 polyunsaturated fatty acid (fish oil) supplementation and the prevention of clinical cardiovascular disease: A science advisory from the American Heart Association. Circulation. 2017; doi:10.1161/CIR.0000000000000482.
• Barley. Natural Medicines. https://naturalmedicines.therapeuticresearch.com. Accessed March 24, 2022.
• Black psyllium. Natural Medicines. https://naturalmedicines.therapeuticresearch.com. Accessed March 24, 2022.
• Nimmagadda R. Allscripts EPSi. Mayo Clinic. April 10, 2024.
• Liao KP. Coronary artery disease in rheumatoid arthritis: Pathogenesis, risk factors, clinical manifestations, and diagnostic implications. https://www.uptodate.com/contents/search. Accessed March 8, 2022.
• What is coronary heart disease? National Heart, Lung, and Blood Institute. https://www.nhlbi.nih.gov/health-topics/coronary-heart-disease Accessed March 8, 2022.
• Kannam JP, et al. Chronic coronary syndrome: Overview of care. https://www.uptodate.com/contents/search. Accessed March 8, 2022.
• Arnett DK, et al. 2019 ACC/AHA guideline on the primary prevention of cardiovascular disease: A report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines. Circulation. 2019; doi:10.1161/CIR.0000000000000678.
• Aspirin use to prevent cardiovascular disease: Preventive medication. U.S. Preventive Services Task Force. https://www.uspreventiveservicestaskforce.org/uspstf/draft-recommendation/aspirin-use-to-prevent-cardiovascular-disease-preventive-medication. Accessed March 23, 2021.
• Zheng SL, et al. Association of aspirin use for primary prevention with cardiovascular events and bleeding events: A systematic review and meta-analysis. JAMA. 2019; doi:10.1001/jama.2018.20578.
• Cutlip D, et al. Revascularization in patients with stable coronary artery disease: Coronary artery bypass graft surgery versus percutaneous coronary intervention. https://www.uptodate.com/contents/search. Accessed March 24, 2022.
• Hypertension in Adults: Screening. U.S. Preventive Services Task Force. https://www.uspreventiveservicestaskforce.org/uspstf/recommendation/hypertension-in-adults-screening. Accessed March 24, 2022.
• How and when to have your cholesterol checked. U.S. Centers for Disease Control and Prevention. https://www.cdc.gov/cholesterol/checked.htm. Accessed March 24, 2022.
• Blond psyllium. Natural Medicines. https://naturalmedicines.therapeuticresearch.com. Accessed March 24, 2022.
• Oats. Natural Medicines. https://naturalmedicines.therapeuticresearch.com. Accessed March 24, 2022.
• Garlic. Natural Medicines. https://naturalmedicines.therapeuticresearch.com. Accessed March 24, 2022.
• Plant sterols. Natural Medicines. https://naturalmedicines.therapeuticresearch.com. Accessed March 24, 2022.
• Ashraf H, et al. Use of flecainide in stable coronary artery disease: An analysis of its safety in both nonobstructive and obstructive coronary artery disease. American Journal of Cardiovascular Drugs. 2021; doi:10.1007/s40256-021-00483-9.
• Ono M, et al. 10-year follow-up after revascularization in elderly patients with complex coronary artery disease. Journal of the American College of Cardiology. 2021; doi:10.1016/j.jacc.2021.04.016.
• Coyle M, et al. A critical review of chronic kidney disease as a risk factor for coronary artery disease. International Journal of Cardiology: Heart & Vasculature. 2021; doi:10.1016/j.ijcha.2021.100822.
• Mankad R (expert opinion). Mayo Clinic. May 9, 2024.
• Scientific Report of the 2020 Dietary Guidelines Advisory Committee. Alcoholic beverages. U.S. Department of Health and Human Services and U.S. Department of Agriculture. https://www.dietaryguidelines.gov/2020-advisory-committee-report. Accessed Feb. 1, 2024.
• Heart disease in women. American Heart Association. https://www.heart.org/en/health-topics/heart-attack/warning-signs-of-a-heart-attack/heart-attack-symptoms-in-women. May 8, 2024.
• Angina treatment: Stents, drugs, lifestyle changes — What's best?
• Coronary artery disease FAQs
• Coronary artery disease: Angioplasty or bypass surgery?
• Coronary artery stent
• Drug-eluting stents
• Four Steps to Heart Health

Associated Procedures

• Cardiac catheterization
• Coronary angiogram
• Coronary angioplasty and stents
• Coronary artery bypass surgery
• Coronary calcium scan
• Echocardiogram
• Electrocardiogram (ECG or EKG)
• Stress test

News from Mayo Clinic

• Mayo Clinic Minute: Signs of coronary artery disease, how to reduce your risk Jan. 24, 2023, 04:15 p.m. CDT
• A History of Cancer, Coronary Artery Disease May Reduce Risk of Dementia Oct. 24, 2022, 02:45 p.m. CDT

Mayo Clinic in Rochester, Minnesota, has been recognized as one of the top Cardiology & Heart Surgery hospitals in the nation for 2023-2024 by U.S. News & World Report.

• Symptoms & causes
• Diagnosis & treatment
• Doctors & departments
• Care at Mayo Clinic

Mayo Clinic does not endorse companies or products. Advertising revenue supports our not-for-profit mission.

• Opportunities

Mayo Clinic Press

Check out these best-sellers and special offers on books and newsletters from Mayo Clinic Press .

• Mayo Clinic on Incontinence - Mayo Clinic Press Mayo Clinic on Incontinence
• The Essential Diabetes Book - Mayo Clinic Press The Essential Diabetes Book
• Mayo Clinic on Hearing and Balance - Mayo Clinic Press Mayo Clinic on Hearing and Balance
• FREE Mayo Clinic Diet Assessment - Mayo Clinic Press FREE Mayo Clinic Diet Assessment
• Mayo Clinic Health Letter - FREE book - Mayo Clinic Press Mayo Clinic Health Letter - FREE book

Your gift holds great power – donate today!

Make your tax-deductible gift and be part of the cutting-edge research and care that's changing medicine.

heart health center / heart a-z list / heart disease (coronary artery disease) article

Heart Disease (Coronary Artery Disease)

• Medical Author: Benjamin Wedro, MD, FACEP, FAAEM
• Medical Editor: Charles Patrick Davis, MD, PhD

What is heart disease (coronary artery disease)?

What causes heart disease, what are the warning signs and symptoms of heart disease, diagnosis of heart disease, what is the treatment for heart disease, can heart disease be prevented, what is the prognosis for heart disease how many people have heart disease.

The heart is like any other muscle in the body. It needs an adequate blood supply to provide oxygen so that the muscle can contract and pump blood to the rest of the body. Not only does the heart pump blood to the rest of the body, but it also pumps blood to itself via the coronary arteries. These arteries originate from the base of the aorta (the major blood vessel that carries oxygenated blood from the heart) and then branch out along the surface of the heart.

When one or more coronary arteries narrow, it may make it difficult for adequate blood to reach the heart, especially during exercise . This can cause the heart muscle to ache like any other muscle in the body. Should the arteries continue to narrow, it may take less activity to stress the heart and provoke symptoms. The classic symptoms of chest pain or pressure and shortness of breath that often spreads to the shoulders, arms, and/or neck due to atherosclerotic heart disease (ASHD) or coronary artery disease (CAD ) are called angina .

Should one of the coronary arteries become completely blocked -- usually due to a plaque that ruptures and causes a blood clot to form -- blood supply to part of the heart may be lost. This causes a piece of the heart muscle to die. This is called a heart attack or myocardial infarction (myo=muscle + cardia=heart + infarction= tissue death).

Cardiovascular disease , for this article, will be limited to describing the spectrum of atherosclerosis or hardening of the arteries that ranges from a minimal blockage that may produce no symptoms to complete obstruction that presents as a myocardial infarction. Other topics, such as myocarditis , heart valve problems, and congenital heart defects will not be covered.

Heart or cardiovascular disease is the leading cause of death in the United States and often can be attributed to lifestyle factors that increase the risk of atherosclerosis or narrowing of arteries. Smoking , along with poorly controlled hypertension ( high blood pressure ), and diabetes , causes inflammation and irritation of the inner lining of the coronary arteries. Over time, cholesterol in the bloodstream can collect in the inflamed areas and begin the formation of plaque. This plaque can grow and as it does, the diameter of the artery narrows. If the artery narrows by 40% to 50%, blood flow is decreased enough to potentially cause the symptoms of angina.

In some circumstances, the plaque can rupture or break open, leading to the formation of a blood clot in the coronary artery. This clot can completely occlude or block the artery. This prevents oxygen-rich blood from being delivered to the heart muscle beyond that blockage and that part of the heart muscle begins to die. This is a myocardial infarction or heart attack. If the situation is not recognized and treated quickly, the affected part of the muscle cannot be revived. It dies and is replaced by scar tissue. Long term, this scar tissue decreases the heart's ability to pump effectively and efficiently and may lead to ischemic cardiomyopathy (ischemic=decreased blood supply + cardio=heart + myo=muscle + pathy=disease).

Heart muscle that lacks adequate blood supply also becomes irritable and may not conduct electrical impulses normally. This can lead to abnormal electrical heart rhythms including ventricular tachycardia and ventricular fibrillation . These are the heart arrhythmias associated with sudden cardiac death.

Who is at risk for heart disease?

There are risk factors that increase the potential to develop plaque within coronary arteries and cause them to narrow. Atherosclerosis (athero=fatty plaque + sclerosis=hardening) is the term that describes this condition. Factors that put people at increa sed risk for heart disease are:

• High blood pressure (hypertension )
• High cholesterol
• Family history of heart problems, especially heart attacks and strokes

Since cardiovascular disease, peripheral artery disease , and stroke share the same risk factors, a patient who is diagnosed with one of the three has increased risk of having or developing the others.

• The classic symptoms of angina, or pain from the heart, are described as a crushing pain or heaviness in the center of the chest with radiation of the pain to the arm (usually the left) or jaw. There can be associated shortness of breath sweating and nausea .
• The symptoms tend to be brought on by activity and get better with rest.
• Some people may have indigestion and nausea while others may have upper abdominal, shoulder, or back pain .
• Unstable angina is the term used to describe symptoms that occur at rest, waken the patient from sleep , and do not respond quickly to nitroglycerin or rest.

Other heart (cardiovascular) disease symptoms and signs

Not all pain from heart disease have the same signs and symptoms. The more we learn about heart disease, the more we realize that symptoms can be markedly different in different groups of people. Women , people who have diabetes, and the elderly may have different pain perceptions and may complain of overwhelming fatigue and weakness or a change in their ability to perform routine daily activities like walking , climbing steps, or doing household chores. Some patients may have no discomfort at all.

Most often, the symptoms of cardiovascular disease become worse over time, as the narrowing of the affected coronary artery progresses over time and blood flow to that part of the heart decreases. It may take less activity to cause symptoms to occur and it may take longer for those symptoms to get better with rest. This change in exercise tolerance is helpful in making the diagnosis.

Often the first signs and symptoms of heart disease may be a heart attack. This can lead to crushing chest pressure, sho rtness of breath, sweating, and perhaps sudden cardiac death.

Health News

• Another Study Finds CRISPR Gene Therapy Fights Sickle Cell
• Is Mom's Fish Intake During Pregnancy Linked to Juvenile Arthritis?
• Pandemic-Era Tax Credits Made Healthcare More Affordable, But They're Set to Expire
• Most Americans Believe Pandemic Policies Were Good Idea: Poll
• High Out-of-Pocket Costs Keep Some From Lifesaving Opioid Antidote

Primary care practitioners, including those in family medicine, internal medicine specialists, and women's health, often help make the initial diagnosis of heart disease and can manage stable patients who do not need invasive procedures or interventions. These providers also work to help minimize potential risk factors for heart disease, so that it does not develop, or if it is already present, to minimize the progression of the artery blockages.

Emergency physicians often make the diagnosis of angina when a patient presents with symptoms of heart disease. As well, when patients present to the ER with symptoms of a heart attack, they work as a team with the cardiologists to treat the patient quickly to restore blood supply to the dying heart muscle.

Cardiologists are specialists who treat cardiovascular heart disease. In addition to confirming the diagnosis using a heart catheterization, they often can perform angioplasty to dilate or open a narrowed or blocked artery and restore blood supply to the heart muscle. As well, cardiologists help manage patients with chest pain to minimize the risk of future heart muscle damage.

Cardiothoracic surgeons operate on the heart and perform coronary artery bypass surgery if the coronary arteries are severely blocked and the patient is not a candidate to have angioplasty. These surgeons also repair or replace heart valves and perform other surgeries that involve the structure of the heart.

The diagnosis of cardiovascular disease begins by taking the patient's history. The health-care professional needs to understand the patient's symptoms and this may be difficult. Often, health-care professionals ask about chest pain, but the patient may deny having pain because they perceive their symptoms as pressure or heaviness. Words also may have different meanings for different people. The patient may describe their discomfort as sharp, meaning intense, while the health-care professional may understand that term to mean stabbing. For that reason, it is important for the patient to be allowed to take the time to describe the symptoms in their own words and have the health-care professional try to clarify the terms being used.

The health-care professional may ask questions about the quality and quantity of pain, where it is located, and where it might travel or radiate. It is important to know about the associated symptoms including shortness of breath, sweating, nausea, vomiting , and indigestion, as well as malaise or fatigue .

The circumstances surrounding the symptoms are also important.

• Are the symptoms brought on by activity?
• Do they get better with rest?
• Since symptoms began, does less activity provoke onset of the symptoms?
• Do the symptoms wake the patient?

These are questions that may help decide wither the angina is stable, progressing, or becoming unstable.

• With stable angina, the activity that is required to initiate the symptoms does not fluctuate. For example, a patient may state that their symptoms are brought on by climbing up two flights of stairs or walking one mile.
• Progressive angina would find the patient stating that the symptoms are brought on by less activity than previously.
• In the case of unstable angina, symptoms may arise at rest or wake the patient from sleep .

Risk factors for cardiovascular disease should be assessed including the prese nce of high blood pressure, diabetes, high cholesterol, smoking history, and family history of cardiovascular disease. A past history of stroke or peripheral artery disease are also important risk factors to be assessed.

Physical examination may not necessarily help make the diagnosis of heart disease, but it can help decide whether other underlying medical problems may be the cause of the patient's symptoms.

There are some clues on physical exam that suggest the presence of narrowed arteries to the heart and coronary artery disease , for example, they the doctor may: Check for high blood pressure. Palpate. (feel) for the pulses in the wrists and feet to see if they are present, and if they are normal in their amplitude and force. Lack of pulses may signal a narrowed or blocked artery in the arm or leg. If one artery is narrowed, perhaps others, like the coronary arteries in the heart, also may be narrowed Auscultating or listening to the neck, abdomen and groin for bruits. A bruit is the sound produced within a narrowed artery due to turbulence caused when decreased blood flow occurs across the narrowed area. Bruits can be heard easily with a stethoscope in the he carotid artery in the neck, the abdominal aorta, and the femoral artery .Check sensation in the feet for numbness, decreased sensation, and peripheral neuropathy .

Moreover, many other important conditions may need to be considered as the cause of symptoms. Examples include those arising from the lung (pulmonary embolus), the aorta ( aortic dissection ), the esophagus ( GERD ), and the abdomen ( peptic ulcer disease, gallbladder disease ).

After the history and physical examination are complete, the health-care professional may require more testing if heart disease is considered a potential diagnosis. There are different ways to evaluate the heart anatomy and function; the type and timing of a test needs to be individualized to each patient and their situation.

Most often, the health-care professional, perhaps in consultation with a cardiologist, will order the least invasive test possible to determine whether coronary artery disease is present. Although heart catheterization is the gold standard to define the anatomy of the heart and to confirm heart disease diagnosis (either with partial or complete blockage or no blockage), this is an invasive test and not necessarily indicated for many patients.

Electrocardiogram ( EKG , ECG )

The heart is an electrical pump and electrodes on the skin can capture and record the impulses generated as electricity travels throughout the heart muscle. Heart muscle that has decreased blood supply conducts electricity differently than normal muscle and these changes can be seen on the EKG.

A normal EKG does not exclude cardiovascular disease and coronary artery blockage; there may be narrowing of the coronary arteries that has yet to cause heart muscle damage. An abnormal EKG may be a "normal" variant for that patient and the result has to be interpreted based upon the patient's circumstances.

If possible, an EKG should be compared to previous tracings looking for changes in the electrical conduction patterns.

Stress tests

It would make sense that during exercise, the heart is asked to work harder and if the heart could be monitored and evaluated during that exercise, a test might uncover abnormalities in heart function . That exercise may occur by asking the patient to walk on a treadmill or ride a bicycle while at the same time, an electrocardiogram is being performed. Medications (adenosine, persantine , dobutamine) can be used to stimulate the heart if the patient is unable to exercise because of poor conditioning, injury, or because of an underlying medical condition.

Echocardiography

Ultrasound examination of the heart to evaluate the anatomy of the heart valves, the muscle, and its function may be performed by a cardiologist. This test may be ordered alone or it may be combined with a stress test to look at heart function during exercise.

Nuclear imaging

A radioactive tracer that is injected into a vein can be used to indirectly assess blood flow to the heart. Technetium or thallium can be injected while a radioactive counter is used to map out how heart muscle cells absorb the radioactive chemical and how it is distributed in heart muscle cells may help determine indirectly whether a blockage exists. An area of the heart with no uptake would suggest that the area is not getting enough blood supply. This test may also be combined with an exercise test.

Cardiac computerized tomography (CT) and magnetic resonance imaging ( MRI )

Using these scans, the anatomy of the coronary arteries can be evaluated, including how much calcium is present in the artery walls and whether blockage or artery narrowing are present. Each test has its benefits and limitations and the risks and benefits of considering a CT or MRI depends upon a patient's situation.

Cardiac catheterization

Cardiac catheterization is the gold standard for coronary artery testing. A cardiologist threads a thin tube through an artery in the groin, elbow, or wrist into the coronary arteries. Dye is injected to assess the anatomy and whether blockages are present. This is called a coronary angiogram .

If a blockage exists, it is possible that angioplasty may be performed. Using the same technique as the angiogram , a balloon is positioned at the site of the obstructing plaque. When the balloon is inflated, the plaque is squashed into the wall of the artery to re-establish blood flow. A stent may then be placed across the previously narrowed segment of artery to prevent it from narrowing again.

The goal of treating cardiovascular disease is to maximize the patient's quantity and quality of life.

• Prevention is the key to avoid cardiovascular disease and optimize treatment.
• Once plaque formation has begun, it is possible to limit its progression by maintaining a healthy lifestyle with routine exercise, diet , and by aiming for lifetime control of high blood pressure, high cholesterol, and diabetes.

Medical Treatment

• Aspirin may be used for its antiplatelet activity, making platelets (one type of blood cell that helps blood clot) less sticky and decreasing the risk of a heart attack. The decision to use aspirin routinely depends upon whether other risk factors for heart disease are present.
• Medications may be prescribed in patients with heart disease to maximize blood flow to the heart and increase the efficiency of the pumping function of the heart.
• Beta-blocker medications help block the action of adrenaline on the heart, slowing the heart rate. These medications also help the heart beat more efficiently and decrease the oxygen requirements of the heart muscle during work.
• Calcium channel blockers also help the heart muscle contract and pump more efficiently.
• Nitrates help dilate arteries and increase blood flow to the heart muscle. They may be short-acting ( Nitrostat ) to treat acute angina symptoms or long-acting preparations (Imdur) may be prescribed for prevention.
• Should there be significant stenosis or narrowing of the coronary arteries, angioplasty and/or stenting (described above) may be considered to open the blocked areas. These procedures are performed in conjunction with cardiac catheterization. Depending upon the patient's anatomy and the extent of the blockage present, coronary artery bypass graft surgery ( CABG ) may be required.
• If a stent is placed, other antiplatelet medications like clopidogrel ( Plavix ) may be prescribed.

Subscribe to MedicineNet's Heart Health Newsletter

By clicking "Submit," I agree to the MedicineNet Terms and Conditions and Privacy Policy . I also agree to receive emails from MedicineNet and I understand that I may opt out of MedicineNet subscriptions at any time.

It may take 10 to 15 years from the beginning of a plaque formation in a coronary artery to narrow that artery to constrict blood flow.

The American Heart Association and the American College of Cardiology have developed guidelines so that healthcare professionals may counsel and treat their patients to decrease the risk of developing heart disease. New attention is being paid to the role of weight reduction, diet , exercise, and the use of cholesterol-lowering medications called statins .

In the past, the goal for statin drugs like atorvastatin was to lower the blood cholesterol level to a specific number, and statins were prescribed for patients with high cholesterol levels or those who had had heart attacks. The new guidelines recommend that more patients may benefit from these statin drugs. Rather than having specific cholesterol numbers as a goal, the new goal is to lower the blood cholesterol level by 50% in high-risk patients and by 30% to 50% in those who are at lower risk to develop heart disease. You and your doctor should discuss which goals are best for your condition.

For patients with a history of heart attack, transient ischemic attack ( TIA ), or stroke, statins may be appropriate for patients with high LDL cholesterol levels (the “bad” cholesterol), for example, those who have type 2 diabetes , and those who have a 10-year risk of a heart attack greater than 7.5%. You and your healthcare professional may estimate risk by using the American Heart Association and American College of Cardiology's ASCVD (Atherosclerotic Cardiovascular Disease) Risk Calculator.

Preventing cardiovascular disease is a lifelong commitment to controlling blood pressure, and high cholesterol, quitting smoking, and diabetes. Now, new opportunities exist to decrease risk even further with these new guidelines. These are also the steps to take to d ecrease the risk of stroke and peripheral artery disease.

What lifestyle changes can a person make to prevent heart disease and heart attacks?

The goal of treating cardiovascular disease is to maximize the person's quantity and quality of life. Prevention is the key to avoiding cardiovascular disease and optimizing treatment. Once plaque formation has begun, it is possible to limit its progression by making these lifestyle changes:

• Maintain a healthy lifestyle with routine exercise
• Quit smoking
• Eat a heart-healthy diet such as the Mediterranean Diet .
• Aim for lifetime control of high blood pressure, high cholesterol, and diabetes.
• An estimated 15.5 million people in the United States have coronary artery disease.
• Each year, 1.5 million patients suffer an acute myocardial infarction and more than 600,000 people die.
• With a better understanding of the different signs and symptoms of heart disease, especially the "atypical" symptoms experienced by women and the elderly, the diagnosis of heart disease has improved.
• The prognosis for the patient is better when diagnosis and treatment are initiated early.
• Educating the public about early access to emergency services when a patient develops acute chest pain can help save lives.
• The goal of the treatment of heart disease is to maximize longevity and quality of life.

Healthy Heart Resources

• Signs Psoriasis May Be Affecting Your Heart
• Diagnosing Hypertrophic Cardiomyopathy
• Exercising With HCM: How to Sweat Safely

Featured Centers

• What Are the Best PsA Treatments for You?
• Understanding Biologics
• 10 Things People With Depression Wish You Knew

Top Heart Disease (Coronary Artery Disease) Related Articles

Heart Healthy Diet: 25 Foods You Should Eat

Living With A-Fib

Chest pain is a common complaint by a patient in the ER. Causes of chest pain include broken or bruised ribs, pleurisy, pneumothorax, shingles, pneumonia, pulmonary embolism, angina, heart attack, costochondritis, pericarditis, aorta or aortic dissection, and reflux esophagitis.

Diagnosis and treatment of chest pain depends upon the cause and clinical presentation of the patient's chest pain.

CT Scan vs. MRI

Nutrition Quiz

Heart Detail Picture

Heart Disease: Causes of a Heart Attack

Heart Disease Quiz

High Blood Pressure (Hypertension)

Cholesterol Quiz

High-Sensitivity Troponin Test

How Long Can You Live With an Implanted Defibrillator?

Cough: 19 Tips on How to Stop a Cough

Parathyroidectomy

Benefits of Sex

What Are The Four Heart Sounds?

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings
• My Bibliography
• Collections
• Citation manager

Save citation to file

Email citation, add to collections.

• Create a new collection
• Add to an existing collection

Add to My Bibliography

Your saved search, create a file for external citation management software, your rss feed.

• Search in PubMed
• Search in NLM Catalog
• Add to Search

Homocysteine concentration in coronary artery disease and severity of coronary lesions

Affiliation.

• 1 Department of Cardiology, Suining Central Hospital, Suining, Sichuan, China.
• PMID: 38896027
• DOI: 10.1111/jcmm.18474

Our previous study reckons that the impact of the rs1801133 variant of 5,10-methylenetetrahydrofolate reductase (MTHFR) on coronary artery disease (CAD) is possibly mediated by cardiometabolic disorder. This study is performed to verify this hypothesis. Four hundred and thirty CAD patients and 216 CAD-free individuals were enrolled in this case-control study. The rs1801133 variant was genotyped by PCR-RFLP. Severity of coronary lesions was evaluated by number of stenotic coronary vessels and extent of coronary stenosis. The rs1801133 T allele significantly increased homocysteine levels in patients with CAD and CAD-free individuals. Individuals with the T allele of rs1801133 had an increased risk of developing CAD. In contrast, individuals with the TT genotype of rs1801133 were at high risk of multiple vessel lesions. The carriers of CT genotype had higher levels of systolic blood pressure (SBP), low-density lipoprotein cholesterol (LDL-C), and high-sensitivity C-reactive protein (hs-CRP), and lower levels of apolipoprotein A1 (APOA1) than those with CC genotype in male patients with CAD. The receiver operating characteristic (ROC) curve and precision-recall (PR) curve indicated that hyperhomocysteinemia was sensitive to predict the severity of CAD. Multivariate logistic regression revealed that homocysteine, rs1801133, age, smoking, weight, body mass index (BMI), lipoprotein(a) [Lp(a)], and hs-CRP were independent risk factors for CAD. The increased risk of CAD and severity of coronary lesions associated with rs1801133 in the Chinese Han population were attributed, at least partly, to high homocysteine levels. Hyperhomocysteinemia had a high predictive value for severe CAD or multiple vessel lesions.

Keywords: 5,10‐methylenetetrahydrofolate reductase; coronary artery disease; homocysteine; multiple vessel lesions.

© 2024 The Author(s). Journal of Cellular and Molecular Medicine published by Foundation for Cellular and Molecular Medicine and John Wiley & Sons Ltd.

PubMed Disclaimer

• Miller AL. The methylation, neurotransmitter, and antioxidant connections between folate and depression. Altern Med Rev. 2008;13:216‐226.
• Wang L, Shangguan S, Chang S, et al. Determining the association between methylenetetrahydrofolate reductase (MTHFR) gene polymorphisms and genomic DNA methylation level: a meta‐analysis. Birth Defects Res A Clin Mol Teratol. 2016;106:667‐674.
• Xhemalce B. From histones to RNA: role of methylation in cancer. Brief Funct Genomics. 2013;12:244‐253.
• Lanouette S, Mongeon V, Figeys D, Couture JF. The functional diversity of protein lysine methylation. Mol Syst Biol. 2014;10:724.
• Weisberg I, Tran P, Christensen B, Sibani S, Rozen R. A second genetic polymorphism in methylenetetrahydrofolate reductase (MTHFR) associated with decreased enzyme activity. Mol Genet Metab. 1998;64:169‐172.
• Search in MeSH
• Citation Manager

NCBI Literature Resources

MeSH PMC Bookshelf Disclaimer

The PubMed wordmark and PubMed logo are registered trademarks of the U.S. Department of Health and Human Services (HHS). Unauthorized use of these marks is strictly prohibited.

• Introduction
• Conclusions
• Article Information

CT indicates computed tomography. Participant numbers in the main study cohort based on adverse pregnancy outcomes (not mutually exclusive categories): any adverse pregnancy outcome, n = 1991; gestational diabetes, n = 144; gestational hypertension, n = 252; preeclampsia, n = 499; preterm delivery, n = 999; small-for-gestational-age infant, n = 526; no adverse pregnancy outcome, n = 8537. The SCORE2 algorithm predicts fatal and nonfatal cardiovascular disease and was developed using data on age, sex, smoking, diabetes, systolic blood pressure, and total and high-density lipoprotein cholesterol. For individuals aged 50 to 69 years, the 10-year cardiovascular risk is classified as follows: less than 5% = low risk; 5% to less than 10% = intermediate risk; 10% or higher = high risk. Among 179 participants with missing SCORE2 data, 164 were missing 1 variable and 15 were missing 2 variables.

The crude model included the adverse pregnancy outcome; the confounder model included the adverse pregnancy outcome, age, smoking, body mass index (BMI) at age 20 years, educational level, financial strain, family history of myocardial infarction and stroke, and total number of deliveries; the predictor model included the adverse pregnancy outcome, age, smoking, diabetes, systolic blood pressure, hypertension and lipid-lowering medications, BMI, and high-density lipoprotein and total cholesterol. In panel B, women without gestational diabetes but with any of gestational hypertension, preeclampsia, preterm delivery, or having a small-for-gestational-age infant contributed to the multivariable models, but their estimates are not shown.

a Total numbers are different for coronary artery calcium score because of missing outcome data.

See Figure 2 legend for description of crude, confounder, and predictor models. In panel A, women without gestational hypertension but with any of gestational diabetes, preeclampsia, preterm delivery, or having a small-for-gestational-age infant contributed to the multivariable models, but their estimates are not shown; in panel B, women without preeclampsia but with any of gestational diabetes, gestational hypertension, preterm delivery, or having a small-for-gestational-age infant contributed to the multivariable models, but their estimates are not shown.

See Figure 2 legend for descriptions of crude, confounder, and predictor models. In panel A, women without preterm delivery but with any of gestational diabetes, gestational hypertension, preeclampsia, or having a small-for-gestational-age infant contributed to the multivariable models, but their estimates are not shown; in panel B, women without a small-for-gestational-age infant but with any of gestational diabetes, gestational hypertension, preeclampsia, or preterm delivery contributed to the multivariable models, but their estimates are not shown.

eAppendix. Supplemental Methods

eReferences

eTable 1. Coronary Artery Disease Prevalence Differences as Vascular Age in Years by Adverse Pregnancy Outcome History

eTable 2. Overall Distribution and Prevalence of Segment-Specific Coronary Atherosclerosis by Adverse Pregnancy Outcome History

eTable 3. Subclinical Coronary Artery Disease in Women With Intermediate 10-Year Cardiovascular Disease Risk by Adverse Pregnancy Outcome History

eTable 4. Associations Between Adverse Pregnancy Outcome and Coronary Artery Disease Indices in Combined Model

eTable 5. Associations Between Preeclampsia History Ascertained by Self-Report and Coronary Artery Disease Indices

eTable 6. Dose Response Association Between Adverse Pregnancy Outcome and Coronary Artery Disease

eTable 7. Association Between Adverse Pregnancy Outcomes and Coronary Artery Disease Indices in Postmenopausal Women

eTable 8. Extended Coronary Artery Calcium Score Data by Adverse Pregnancy Outcome History

Data Sharing Statement

• Adverse Pregnancy Outcomes—Risk Enhancers Whose Time Has Finally Arrived JAMA Editorial February 7, 2023 Natalie A. Bello, MD, MPH

See More About

Select your interests.

Customize your JAMA Network experience by selecting one or more topics from the list below.

• Academic Medicine
• Acid Base, Electrolytes, Fluids
• Allergy and Clinical Immunology
• American Indian or Alaska Natives
• Anesthesiology
• Anticoagulation
• Art and Images in Psychiatry
• Artificial Intelligence
• Assisted Reproduction
• Bleeding and Transfusion
• Caring for the Critically Ill Patient
• Challenges in Clinical Electrocardiography
• Climate and Health
• Climate Change
• Clinical Challenge
• Clinical Decision Support
• Clinical Implications of Basic Neuroscience
• Clinical Pharmacy and Pharmacology
• Complementary and Alternative Medicine
• Consensus Statements
• Coronavirus (COVID-19)
• Critical Care Medicine
• Cultural Competency
• Dental Medicine
• Dermatology
• Diabetes and Endocrinology
• Diagnostic Test Interpretation
• Drug Development
• Electronic Health Records
• Emergency Medicine
• End of Life, Hospice, Palliative Care
• Environmental Health
• Equity, Diversity, and Inclusion
• Facial Plastic Surgery
• Gastroenterology and Hepatology
• Genetics and Genomics
• Genomics and Precision Health
• Global Health
• Guide to Statistics and Methods
• Hair Disorders
• Health Care Delivery Models
• Health Care Economics, Insurance, Payment
• Health Care Quality
• Health Care Reform
• Health Care Safety
• Health Care Workforce
• Health Disparities
• Health Inequities
• Health Policy
• Health Systems Science
• History of Medicine
• Hypertension
• Images in Neurology
• Implementation Science
• Infectious Diseases
• Innovations in Health Care Delivery
• JAMA Infographic
• Law and Medicine
• Leading Change
• Less is More
• LGBTQIA Medicine
• Lifestyle Behaviors
• Medical Coding
• Medical Devices and Equipment
• Medical Education
• Medical Education and Training
• Medical Journals and Publishing
• Mobile Health and Telemedicine
• Narrative Medicine
• Neuroscience and Psychiatry
• Notable Notes
• Nutrition, Obesity, Exercise
• Obstetrics and Gynecology
• Occupational Health
• Ophthalmology
• Orthopedics
• Otolaryngology
• Pain Medicine
• Palliative Care
• Pathology and Laboratory Medicine
• Patient Care
• Patient Information
• Performance Improvement
• Performance Measures
• Perioperative Care and Consultation
• Pharmacoeconomics
• Pharmacoepidemiology
• Pharmacogenetics
• Pharmacy and Clinical Pharmacology
• Physical Medicine and Rehabilitation
• Physical Therapy
• Physician Leadership
• Population Health
• Primary Care
• Professional Well-being
• Professionalism
• Psychiatry and Behavioral Health
• Public Health
• Pulmonary Medicine
• Regulatory Agencies
• Reproductive Health
• Research, Methods, Statistics
• Resuscitation
• Rheumatology
• Risk Management
• Scientific Discovery and the Future of Medicine
• Shared Decision Making and Communication
• Sleep Medicine
• Sports Medicine
• Stem Cell Transplantation
• Substance Use and Addiction Medicine
• Surgical Innovation
• Surgical Pearls
• Teachable Moment
• Technology and Finance
• The Art of JAMA
• The Arts and Medicine
• The Rational Clinical Examination
• Tobacco and e-Cigarettes
• Translational Medicine
• Trauma and Injury
• Treatment Adherence
• Ultrasonography
• Users' Guide to the Medical Literature
• Vaccination
• Venous Thromboembolism
• Veterans Health
• Women's Health
• Workflow and Process
• Wound Care, Infection, Healing

Others Also Liked

• Download PDF
• X Facebook More LinkedIn

Sederholm Lawesson S , Swahn E , Pihlsgård M, et al. Association Between History of Adverse Pregnancy Outcomes and Coronary Artery Disease Assessed by Coronary Computed Tomography Angiography. JAMA. 2023;329(5):393–404. doi:10.1001/jama.2022.24093

Manage citations:

© 2024

• Permissions

Association Between History of Adverse Pregnancy Outcomes and Coronary Artery Disease Assessed by Coronary Computed Tomography Angiography

• 1 Department of Cardiology, Linköping University Hospital, and Department of Health, Medicine, and Caring Sciences, Linköping University, Linköping, Sweden
• 2 Perinatal and Cardiovascular Epidemiology, Lund University Diabetes Centre, Department of Clinical Sciences Malmö, Lund University, Malmö, Sweden
• 3 Department of Public Health and Clinical Medicine, Umeå University, Umeå, Sweden
• 4 Department of Cardiology, Sahlgrenska University Hospital, Gothenburg, Sweden
• 5 Department of Molecular and Clinical Medicine, Institute of Medicine, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden
• 6 Department of Clinical Science, Intervention, and Technology, Karolinska Institutet, Stockholm, Sweden
• 7 Department of Radiology, Capio St Görans Hospital, Stockholm, Sweden
• 8 Department of Obstetrics and Gynecology, Linköping University Hospital, and Department of Biomedical and Clinical Sciences, Linköping University, Linköping, Sweden
• 9 Department of Medical Sciences, Cardiology, Uppsala University, Uppsala, Sweden
• 10 Department of Cardiology, Skåne University Hospital, Malmö, Sweden
• 11 Cardiovascular Research Translational Studies, Department of Clinical Sciences Malmö, Lund University, Malmö, Sweden
• 12 Department of Obstetrics and Gynecology, Skåne University Hospital, Lund and Malmö, Sweden
• 13 Department of Clinical Sciences, Danderyd University Hospital, Karolinska Institutet, Stockholm, Sweden
• 14 Unit of Cardiovascular and Nutritional Epidemiology, Institute of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden
• 15 Department of Medicine, Geriatrics, and Emergency Medicine, Sahlgrenska University Hospital, Östra Hospital, Gothenburg, Sweden
• 16 Clinical Epidemiology Division, Department of Medicine, Solna, Karolinska Institutet, Stockholm, Sweden
• 17 Department of Obstetrics and Gynecology, University of Gothenburg, Institute of Clinical Sciences, Sahlgrenska Academy, Gothenburg, Sweden
• 18 Department of Women’s and Children’s Health, Uppsala University, Uppsala, Sweden
• Editorial Adverse Pregnancy Outcomes—Risk Enhancers Whose Time Has Finally Arrived Natalie A. Bello, MD, MPH JAMA

Question   Are there associations between a history of adverse pregnancy outcomes and coronary atherosclerosis independent of cardiovascular risk in women aged 50 to 65 years?

Findings   In this population-based cross-sectional analysis of Swedish women undergoing screening coronary computed tomography (CT) angiography, there was a statistically significant association between history of adverse pregnancy outcomes and image-identified coronary artery disease, including preeclampsia (prevalence, 36.3% vs 28.3%) and gestational hypertension (prevalence, 40.9% vs 28.3%). This association was present in the subgroup of women estimated to be at low cardiovascular disease risk.

Meaning   A history of adverse pregnancy outcomes was significantly associated with coronary CT image–identified coronary artery disease.

Importance   Adverse pregnancy outcomes are recognized risk enhancers for cardiovascular disease, but the prevalence of subclinical coronary atherosclerosis after these conditions is unknown.

Objective   To assess associations between history of adverse pregnancy outcomes and coronary artery disease assessed by coronary computed tomography angiography screening.

Design, Setting, and Participants   Cross-sectional study of a population-based cohort of women in Sweden (n = 10 528) with 1 or more deliveries in 1973 or later, ascertained via the Swedish National Medical Birth Register, who subsequently participated in the Swedish Cardiopulmonary Bioimage Study at age 50 to 65 (median, 57.3) years in 2013-2018. Delivery data were prospectively collected.

Exposures   Adverse pregnancy outcomes, including preeclampsia, gestational hypertension, preterm delivery, small-for-gestational-age infant, and gestational diabetes. The reference category included women with no history of these exposures.

Main Outcomes and Measures   Coronary computed tomography angiography indexes, including any coronary atherosclerosis, significant stenosis, noncalcified plaque, segment involvement score of 4 or greater, and coronary artery calcium score greater than 100.

Results   A median 29.6 (IQR, 25.0-34.9) years after first registered delivery, 18.9% of women had a history of adverse pregnancy outcomes, with specific pregnancy histories ranging from 1.4% (gestational diabetes) to 9.5% (preterm delivery). The prevalence of any coronary atherosclerosis in women with a history of any adverse pregnancy outcome was 32.1% (95% CI, 30.0%-34.2%), which was significantly higher (prevalence difference, 3.8% [95% CI, 1.6%-6.1%]; prevalence ratio, 1.14 [95% CI, 1.06-1.22]) compared with reference women. History of gestational hypertension and preeclampsia were both significantly associated with higher and similar prevalence of all outcome indexes. For preeclampsia, the highest prevalence difference was observed for any coronary atherosclerosis (prevalence difference, 8.0% [95% CI, 3.7%-12.3%]; prevalence ratio, 1.28 [95% CI, 1.14-1.45]), and the highest prevalence ratio was observed for significant stenosis (prevalence difference, 3.1% [95% CI, 1.1%-5.1%]; prevalence ratio, 2.46 [95% CI, 1.65-3.67]). In adjusted models, odds ratios for preeclampsia ranged from 1.31 (95% CI, 1.07-1.61) for any coronary atherosclerosis to 2.21 (95% CI, 1.42-3.44) for significant stenosis. Similar associations were observed for history of preeclampsia or gestational hypertension among women with low predicted cardiovascular risk.

Conclusions and Relevance   Among Swedish women undergoing coronary computed tomography angiography screening, there was a statistically significant association between history of adverse pregnancy outcomes and image-identified coronary artery disease, including among women estimated to be at low cardiovascular disease risk. Further research is needed to understand the clinical importance of these associations.

Coronary artery disease is the most common cardiovascular disease and the leading cause of death, and women and men share most of its modifiable risk factors. 1 , 2 However, some risk factors, such as adverse pregnancy outcomes 3 are unique to women. Women with previous preeclampsia, 4 - 7 gestational hypertension, small for gestational age infant, 8 preterm delivery, 9 or gestational diabetes 10 , 11 all have higher risk of coronary artery disease compared with women without these complications. While some adverse pregnancy outcomes are associated with cardiovascular disease independent of conventional risk factors, pregnancy history does not seem to improve cardiovascular risk prediction models substantially. 12 , 13 Younger women are generally estimated to have low cardiovascular risk, and prevention efforts are often overlooked despite a previous adverse pregnancy outcome. 3 Coronary computed tomography (CT) angiography allows for noninvasive visualization of the coronary arteries and can serve as an effective tool for screening for subclinical coronary artery disease. 14 By elucidating the extent to which the development of coronary atherosclerosis differs by adverse pregnancy outcome history and predicted cardiovascular risk, the association between pregnancy history and coronary artery disease could be better understood. To accomplish this, the purpose of this study was to assess associations between subclinical coronary artery disease assessed by coronary CT angiography screening in women aged 50 to 65 years and history of adverse pregnancy outcomes.

In this cross-sectional study of a population-based cohort, women with 1 or more deliveries in 1973 or later, ascertained via the Swedish National Medical Birth Register, who subsequently at age 50 to 65 years had participated in the Swedish Cardiopulmonary Bioimage Study (SCAPIS) in 2013-2018, were included. Study participants were invited based on their legal sex (male or female) in the Swedish population register at the time of study recruitment. The study was approved by the Swedish Ethical Review Authority (2019-03229) following approval of data collection (2010-228-31M), and all participants provided written informed consent.

In 2013-2018, 30 154 women and men (aged 50-65 years) were recruited at random from the Swedish census population, and during multiple visits underwent cardiovascular examinations and laboratory testing, including completing a questionnaire regarding their health, social circumstances, and lifestyle. 15 , 16 Of the 15 508 women included, 12 825 underwent coronary CT angiography with assessable images of the proximal coronary segments (segments 1, 5, 6, and 11), 15 , 16 and of these, 10 528 women had ascertainable registry data on pregnancy history ( Figure 1 ). For the analysis on subclinical coronary atherosclerosis in women with low (<5%) 10-year cardiovascular risk, we further excluded 502 women with a history of myocardial infarction, coronary revascularization, or stroke and/or missing data for cardiovascular risk prediction variables.

All adverse pregnancy outcomes except gestational diabetes were identified through linkage with the Swedish National Medical Birth Register 1973-2018 (eAppendix in Supplement 1 ), containing data for more than 98% of all deliveries in Sweden. 17 Preeclampsia and gestational hypertension were mainly defined by having received any qualifying International Classification of Diseases (eighth, ninth, or tenth revisions) diagnosis (eAppendix in Supplement 1 ) at any delivery. A small-for-gestational-age infant was defined as a singleton infant with birth weight more than 2 SDs below the mean according to gestational week and sex. 18 Preterm delivery was defined as delivery at 36 weeks 6 days’ gestation or earlier. History of gestational diabetes was self-reported at the time of coronary CT angiography. Each adverse pregnancy outcome was independently defined irrespective of any other adverse pregnancy outcomes occurring during the same or other pregnancies.

The collection of coronary CT angiography and coronary artery calcium data have been previously reported. 15 , 16 In summary, noncontrast coronary artery calcium images were obtained using electrocardiogram (ECG)-gated CT imaging at 120 kV. Coronary CT angiography was performed with a dual-source CT scanner equipped with a Stellar detector (Somatom Definition Flash, Siemens Medical Solutions), with iohexol, 325 mg iodine per kilogram, used as a contrast medium (350 mg iodine per milliliter; GE HealthCare). For coronary assessment, the 18-coronary-segment model was used, with segments 1 through 3, 5 through 7, 9, 11 through 13, and 17 compulsorily reported. 16 Five complementary coronary atherosclerosis indexes were used as outcomes: (1) any coronary atherosclerosis; (2) any significant stenosis (≥50% lumen obstruction); (3) any noncalcified plaque; (4) segment involvement score of 4 or greater (indicating elevated risk of cardiovascular events) 19 ; and (5) coronary artery calcium score of greater than 100 Agatston units (indicating highly calcified atherosclerosis). 20

The data collected at the time of coronary CT angiography were used to describe the study sample, as covariables in regression modeling and for calculation of 10-year cardiovascular risk according to the SCORE2 algorithm (eAppendix in Supplement 1 ). 15 Using data on age, sex, smoking, diabetes, systolic blood pressure, and total and high-density lipoprotein cholesterol, the SCORE2 algorithm was developed, calibrated, and validated to predict the 10-year risk of first-onset fatal and nonfatal cardiovascular disease in European populations aged 40 to 69 years without diabetes or previous cardiovascular disease. 21

Descriptive data are presented as numbers and percentages or medians and interquartile ranges as appropriate.

To study the segment-specific pattern of coronary atherosclerosis by adverse pregnancy outcome history, χ 2 tests were applied to the “any coronary atherosclerosis” counts in 6 coronary artery segments with high known prevalence of coronary atherosclerosis (eAppendix in Supplement 1 ). 16

All comparisons of prevalence were made with women without any adverse pregnancy outcome as the reference. To examine associations between each adverse pregnancy outcome and the coronary CT angiography outcomes, multivariable logistic regression modeling was performed (see the eAppendix in Supplement 1 for further rationale), and odds ratios (ORs) with 95% CIs were calculated. The confounder models included covariables collected at the time of coronary CT angiography that were not postpregnancy predictors or mediators, including age, smoking, reported body mass index at age 20 years, educational level, financial strain, family history of myocardial infarction, family history of stroke, and total number of deliveries. The clinical predictor models included established predictors of coronary atherosclerosis collected at the time of coronary CT angiography: age, smoking, diabetes, systolic blood pressure, hypertension medication, lipid-lowering medication, body mass index, high-density lipoprotein cholesterol, and total cholesterol.

To enhance interpretation, the associations between adverse pregnancy outcome history and coronary CT angiography outcomes were calculated as difference in vascular age, ie, the hypothetical mean age increase that the magnitude of each association corresponds to (eAppendix in Supplement 1 ).

To investigate subclinical coronary atherosclerosis in women with low cardiovascular risk, women with previous cardiovascular disease were excluded. The 10-year cardiovascular risk was calculated using the SCORE2 algorithm. 21 In women with low (<5%) predicted risk, the prevalence of coronary atherosclerosis indexes by pregnancy history was analyzed, as were prevalence differences and prevalence ratios compared with women without previous adverse pregnancy outcomes. Methods of complementary analyses are described in the eAppendix in Supplement 1 .

Multiple imputations of covariables with chained equations were used as needed for the regression models (eAppendix in Supplement 1 ). A 2-sided P  < .05 was set for statistical significance. Because of the potential for type I error due to multiple comparisons, findings should be interpreted as exploratory. All statistical analyses were performed using SAS version 9.4 (SAS Institute Inc).

The main study cohort, described in Table 1 , comprised 10 528 women with at least 1 delivery registered in the Swedish National Medical Birth Register and subsequent assessable coronary CT angiography. Of all women (median age, 57.3 years), 1991 (18.9%) had a history of any adverse pregnancy outcome. The prevalence of each of these was as follows: preeclampsia, 4.7%; gestational hypertension, 2.4%; preterm delivery, 9.5%; small-for-gestational-age infant, 5.0%; and gestational diabetes, 1.4%. Compared with women with no adverse pregnancy outcome history, women with previous preeclampsia or gestational hypertension had numerically higher systolic blood pressure and women with previous gestational diabetes had higher prevalence of diabetes. The median time between first registered delivery and angiography was 29.6 years (range, 0.7-45.5 [IQR, 25.0-34.9] years).

The prevalence, prevalence ratios, and crude and multivariable adjusted ORs of all indexes are presented in Figure 2 , Figure 3 , and Figure 4 . Among women with a history of any adverse pregnancy outcome, 32.1% (95% CI, 30.0%-34.2%) had any coronary atherosclerosis, which was significantly higher (prevalence difference, 3.8% [95% CI, 1.6%-6.1%]; prevalence ratio, 1.14 [95% CI, 1.06-1.22]) compared with women without any history of adverse pregnancy outcomes. History of gestational hypertension and preeclampsia were both significantly associated with higher and similar prevalence of all outcome indexes. For preeclampsia, the highest prevalence difference was observed for any coronary atherosclerosis (prevalence difference, 8.0% [95% CI, 3.7%-12.3%]; prevalence ratio, 1.28 [95% CI, 1.14-1.45]) and the highest prevalence ratio was observed for significant stenosis (prevalence difference, 3.1% [95% CI, 1.1%-5.1%]; prevalence ratio, 2.46 [95% CI, 1.65-3.67]). For preeclampsia, the predictor-adjusted ORs ranged from 1.31 (95% CI, 1.07-1.61) for coronary atherosclerosis to 2.21 (95% CI, 1.42-3.44) for significant stenosis, with similar findings for gestational hypertension (ORs, 1.33 [95% CI, 1.01-1.74] and 1.83 [95% CI, 1.00-3.33], respectively). Women with a previous preterm delivery, small-for-gestational-age infant, or gestational diabetes also had significantly higher prevalences of most indexes, but numerically lower prevalences compared with those with previous preeclampsia or gestational hypertension. In models including clinical predictors, having a small-for-gestational-age infant was significantly associated with a segment involvement score of 4 or greater (OR, 1.51; 95% CI, 1.06-2.14) and significant stenosis (OR, 1.80; 95% CI, 1.13-2.87). Preterm delivery and gestational diabetes were not significantly associated with any of the outcome indexes in the model including clinical predictors, although the 95% CIs for gestational diabetes were wide.

When estimating vascular age by adverse pregnancy outcome, women with a history of preeclampsia or gestational hypertension were 4 years older than those without any history according to the “any coronary atherosclerosis” outcome and 6 to 11 years older according to the more severe indexes (eTable 1 in Supplement 1 ). The corresponding findings for small-for-gestational-age infant, preterm delivery, and gestational diabetes history were more scattered and attenuated.

The distribution of segment-specific coronary atherosclerosis in women with any adverse pregnancy outcome, preeclampsia, gestational hypertension, or small-for-gestational-age infant differed significantly from that of women with no adverse pregnancy outcome (eTable 2 in Supplement 1 ). No significant difference in distribution was observed for preterm delivery or gestational diabetes. The proximal left anterior descending coronary artery was most frequently affected by atherosclerosis, regardless of adverse pregnancy outcome history, and all adverse pregnancy outcomes were typically associated with higher prevalence of coronary atherosclerosis in all studied segments.

In total, 8334 (83.1% of eligible) women were classified as having a low 10-year cardiovascular risk. Among these women, a history of any adverse pregnancy outcome was associated with significantly higher prevalence of all outcomes compared with women with no history of adverse pregnancy outcomes ( Table 2 ). Among the specific adverse pregnancy outcomes, preeclampsia and gestational hypertension were statistically associated with higher prevalence of all coronary indexes, except for the association between gestational hypertension and significant stenosis. The highest prevalence ratio was observed for the association between preeclampsia and significant stenosis (3.15; 95% CI, 1.90-5.21) and the highest prevalence difference between gestational hypertension and any coronary atherosclerosis (10.4%; 95% CI, 3.3%-17.5%). There were no significantly increased prevalence ratios observed for the other adverse pregnancy outcomes. The prevalence of significant stenosis (4.5%) in women with a history of preeclampsia and low predicted cardiovascular risk was numerically on par with that observed in women with no adverse pregnancy outcome history and intermediate (5% to <10%) cardiovascular risk (4.8%; eTable 3 in Supplement 1 ).

Adding the covariables included in the confounder model to the predictor model did not substantially affect model estimates reported in the main analyses (eTable 4 in Supplement 1 ). Instead defining history of preeclampsia through self-report (n = 652 [6.2%]) attenuated model estimates but did not change the overall pattern (eTable 5 in Supplement 1 ). In women with at least 2 registered deliveries, having any adverse pregnancy outcome in more than 1 pregnancy (vs having any adverse pregnancy outcome only once) was significantly associated with a segment involvement score of 4 or greater (OR, 1.63; 95% CI, 1.06-2.51) and a coronary artery calcium score of greater than 100 (OR, 1.67; 95% CI, 1.11-2.52) in predictor-adjusted models (eTable 6 in Supplement 1 ). Limiting the sample to postmenopausal women slightly attenuated estimates overall (eTable 7 in Supplement 1 ). Extended coronary artery calcium score descriptive data by adverse pregnancy outcome history are shown in eTable 8 in Supplement 1 . All adverse pregnancy outcomes were associated with numerically higher median coronary artery calcium scores.

Among Swedish women aged 50 to 65 years undergoing coronary CT angiography screening, there was a statistically significant association between history of adverse pregnancy outcomes and image-identified coronary artery disease, including among women estimated to be at low risk of cardiovascular disease. These associations were primarily found for hypertensive disorders of pregnancy—ie, preeclampsia or gestational hypertension—which were associated with all 5 coronary artery disease indexes. A history of preeclampsia or gestational hypertension was associated with a different distribution of segment-specific coronary atherosclerosis and more than 2 times higher risk of subclinical obstructive and widespread coronary artery disease, which is comparable with studies of incident clinical events. 5 , 22 - 24

Recent observations from the Nurses’ Health Study II suggest that a larger proportion of the associated cardiovascular risk is mediated by postpregnancy risk factors in gestational hypertension compared with preeclampsia. 25 As highlighted by another recent study, women with a history of hypertensive disorders of pregnancy are at increased risk of cardiac remodeling 10 years postpartum due to their higher risk of hypertension. 26 As endothelial dysfunction is impaired in women with a history of preeclampsia 27 or a small-for-gestational-age infant, 28 it might constitute an etiological pathway to understanding atherosclerosis development in women with previous preeclampsia, gestational hypertension, or a small-for-gestational-age infant. It is notable that it is for these 3 complications, which have etiological and clinical overlaps during pregnancy, 29 that the most consistent findings with coronary atherosclerosis in middle age were found.

In contrast, preterm delivery is a much more heterogeneous entity; it can for instance be caused by iatrogenic delivery due to preeclampsia, infection, or premature contractions. While women with a history of preterm delivery, compared with those without, had a higher coronary artery calcium score, no higher prevalence of abnormal coronary CT angiography findings was found. As in the CARDIA study, women with previous gestational diabetes had a higher coronary artery calcium score than those without. 10 The results of the current study support that the risk of coronary atherosclerosis in these women is mainly mediated by conventional risk factors, especially diabetes, 30 as the higher odds of a high calcium score and high atherosclerotic burden were abrogated following multivariable adjustment.

While this study supports previous reports that women with, compared with those without, a history of preeclampsia have higher coronary artery calcium score, 4 , 31 - 33 studies on subclinical coronary artery disease by adverse pregnancy outcome history are scarce. Still, 2 previous European studies have provided important information on the prevalence of coronary atherosclerosis in women with a history of preeclampsia. In the CREW-IMAGO study, 34.5% of women aged 40 to 63 years with a history of preeclampsia had any coronary atherosclerosis and 2.9% had significant stenosis. 32 In the recent Danish CPH-PRECIOUS study, coronary atherosclerosis was detected in 27.4% of women aged 40 to 55 years with previous preeclampsia, compared with 20% in age- and parity-matched controls. 34 These prevalence numbers are in line with those reported herein, given the older mean age of women in the study.

The results of the complementary analyses on a dose-response association of adverse pregnancy outcomes on coronary artery disease indexes warrant detailed investigation in future studies. The attenuated model estimates observed in postmenopausal women are in line with previous findings, suggesting that the relative association between adverse pregnancy outcomes and incident cardiovascular disease decreases over decades. 35

A major strength of the present study is the utilization of a cohort with data on pregnancy history for all women ascertained via a population-based medical birth registry covering 5 decades. Thus, the study was not limited to comparing women with history of an adverse pregnancy outcome with a general sample of parous women with unknown pregnancy outcome status or relying on self-reported data collected many years after delivery, which is associated with recall bias. Another advantage is the investigation of 5 relevant adverse pregnancy outcomes for coronary atherosclerosis, and not only preeclampsia. Furthermore, all women included in the cohort also underwent standardized coronary CT angiography at a time point that was independent of adverse pregnancy outcome status. Altogether, this allows for better estimations of the prevalence of coronary atherosclerosis in women with a history of each respective adverse pregnancy outcome compared with those without, ultimately providing considerably more informative evidence for clinical decisions in the preventive context.

This study has several limitations. First, as is the case with many statistical comparisons, there is a potential for type I error. The results are to some degree exploratory and should be interpreted based on their overall pattern, rather than by any specific finding, as false-positive associations may have been found. Second, as women had to survive until they were aged 50 to 65 years to be included in the study, the focus was on women who are potential targets for prevention efforts at this age and the results should not be interpreted as an effort to estimate causal effects. Third, information on smoking, as well as gestational diabetes at the time of pregnancy, was not available for all women, and was captured at the time of coronary CT angiography. The prevalence of previous gestational diabetes was low in this cohort, but this was expected given the low past prevalence of gestational diabetes in Sweden in areas both with and without universal screening. 36 Fourth, although the Swedish National Medical Birth Register has been shown to be complete and accurate, misclassification may have occurred. Still, an underestimation of associations is most likely if misclassification occurred at random and was not related to outcomes. Fifth, because of the population-based design, low-dose radiation was used for coronary CT angiography, and thus high-quality images could not be obtained from all participants. However, a low-dose protocol is more feasible for clinical use in the prevention setting.

Among Swedish women undergoing screening coronary CT angiography, there was a statistically significant association between history of adverse pregnancy outcomes and image-identified coronary artery disease, including among women estimated to be at low cardiovascular disease risk. Further research is needed to understand the clinical importance of these associations.

Corresponding Authors: Sofia Sederholm Lawesson, MD, PhD, Department of Cardiology, Linköping University Hospital, SE-58185 Linköping, Sweden ( [email protected] ); Simon Timpka, MD, PhD, Perinatal and Cardiovascular Epidemiology, Lund University Diabetes Centre, Box 50332, SE-20213 Malmö, Sweden ( [email protected] ).

Accepted for Publication: December 13, 2022.

Author Contributions: Drs Sederholm Lawesson and Timpka had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Sederholm Lawesson, Swahn, Bergdahl, Gonçalves, Gunnarsson, Jernberg, Pehrson, Rosengren, Wikström, Timpka.

Acquisition, analysis, or interpretation of data: Sederholm Lawesson, Swahn, Pihlsgård, Andersson, Angerås, Bacsovics Brolin, Bergdahl, Blomberg, Christersson, Gonçalves, Gunnarsson, Jernberg, Johnston, Leander, Lilliecreutz, Rosengren, Anette Sandström, Anna Sandström, Sarno, Själander, Svanvik, Thunström, Wikström, Timpka.

Drafting of the manuscript: Sederholm Lawesson, Swahn, Lilliecreutz, Timpka.

Critical revision of the manuscript for important intellectual content: Sederholm Lawesson, Swahn, Pihlsgård, Andersson, Angerås, Bacsovics Brolin, Bergdahl, Blomberg, Christersson, Gonçalves, Gunnarsson, Jernberg, Johnston, Leander, Lilliecreutz, Pehrson, Rosengren, Anette Sandström, Anna Sandström, Sarno, Själander, Svanvik, Thunström, Wikström.

Statistical analysis: Pihlsgård.

Obtained funding: Jernberg, Rosengren, Timpka.

Administrative, technical, or material support: Andersson, Gonçalves, Lilliecreutz, Rosengren, Timpka.

Supervision: Sederholm Lawesson, Swahn, Andersson, Blomberg, Christersson, Gonçalves, Lilliecreutz, Anna Sandström, Svanvik, Timpka.

Conflict of Interest Disclosures: Dr Sederholm Lawesson reported receipt of speaker fees from Bayer and Pfizer. Dr Andersson reported receipt of speaker fees from or advisory board membership in Actelion Pharmaceuticals Ltd, Vifor Pharma, Pharmacosmos, and AstraZeneca. Dr Christersson reported receipt of institutional research grants from Pfizer; speaker fees from or advisory board membership in Bristol Myers Squibb, Boehringer Ingelheim, Bayer, Novartis, Orion Pharma, and AstraZeneca; and personal fees from event adjudication from Uppsala Clinical Research Center. Dr Själander reported receipt of personal fees from Bristol Myers Squibb and Bayer. No other disclosures were reported.

Funding/Support: The main funding body of SCAPIS is the Swedish Heart and Lung Foundation. The study is also funded by the Knut and Alice Wallenberg Foundation, the Swedish Research Council, VINNOVA, the University of Gothenburg and Sahlgrenska University Hospital, Karolinska Institutet and Karolinska University Hospital, Linköping University and University Hospital, Lund University and Skåne University Hospital, Umeå University and University Hospital, and Uppsala University and University Hospital. Support was also received from ALF grant Region Östergötland RÖ-966520 (Dr Sederholm Lawesson); Swedish Research Council Strategic Research Area Exodiab grant Dnr 2009-1039, Linnaeus grant Dnr 349-2006-23, and Swedish Foundation for Strategic Research grant Dnr IRC15-006, the Swedish Heart and Lung Foundation, Skåne University Hospital, and the Lund University Diabetes Center (Dr Gonçalves); Swedish Research Council grant 2018-02527 and AFA Insurance (Dr Rosengren); and Swedish Research Council grant 2019-02082, Swedish Heart and Lung Foundation grant 20180312, the Jeansson Foundation, Åke Wiberg Foundation, and research support from the health care authority in Region Skåne and Lund University (Dr Timpka).

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication.

Data Sharing Statement: See Supplement 2 .

• Register for email alerts with links to free full-text articles
• Access PDFs of free articles
• Manage your interests
• Save searches and receive search alerts

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• BMJ Case Rep

Case Report

Rapidly progressive coronary artery disease as the first manifestation of antiphospholipid syndrome, abdullah sayied abdullah.

1 Department of Cardiology, University Hospital Limerick, Limerick, Ireland

Hatim Yagoub

Thomas j kiernan, caroline daly.

2 Department of Cardiology, St James Hospital, Dublin, Ireland

Antiphospholipid syndrome (APS) is an autoimmune multisystem disorder characterised by high incidence of arterial and venous thrombosis. Cardiovascular manifestations also include valvular heart disease, ventricular thrombi and higher risk for coronary artery disease (CAD). In this case report, we describe a 61-year-old woman who had no significant risk factors for CAD, and presented with aggressive disease in native and graft vessels that required multiple coronary interventions. The extent of her aggressive CAD could not be explained by her risk factors profile. Therefore autoantibodies screening was carried out and showed a strongly positive anticardiolipin and β 2 glycoprotein-I antibody, and hence a diagnosis of antiphospholipid syndrome was made.

Antiphospholipid syndrome (APS) is an autoimmune multisystem disease. It occurs secondary to another autoimmune disorder, but majority of cases are primary. It usually presents with recurrent miscarriages or recurrent thromboembolic disease. Cardiovascular complications include valvular heart disease, but aggressive coronary artery disease is rare. 1 To the best of our knowledge, APS presenting solely with multiple acute coronary syndromes that necessitated coronary artery bypass grafting (CABG) and multiple percutaneous coronary interventions (PCI) to native vessels has not been reported previously.

The condition results from the formation of autoantibodies against phospholipids and β 2 glycoprotein-I. Failure of macrophages to clear apoptotic cell membranes, histones and nuclear material leads to the deposition of debris in the lymphoid tissues and prevents their uptake by antigen presenting cells. T cells recognise these self-antigens and stimulate B-cell response and subsequent autoantibody formation, with deposition of immune complexes in endothelial cells, monocytes, platelets and trophoblasts. This alters the functioning of these cells leading to classical features of APS. 2 3 Persistent positive lupus anticoagulant test and antiphospholipid antibodies are required for diagnosis. 4 These include anticardiolipin and anti β 2 glycoprotein-I antibodies.

Inflammation has long been postulated as a key factor in atherosclerosis, and research is still ongoing to define its relative role. 5 In this report we describe an association of aggressive coronary artery disease with APS.

Case presentation

A 61-year-old woman presented with aggressive coronary artery disease as an unusual first manifestation of antiphospholipid syndrome.

The patient had no risk factors for coronary artery disease (CAD) or history of hormone replacement therapy. She did have one miscarriage in her first pregnancy but subsequent five pregnancies were uneventful. Her initial presentation was with a non-ST segment myocardial infarction (NSTEMI). Urgent coronary angiography demonstrated left main stem (LMS) 70% ostial stenosis, proximal left anterior descending (LAD) stenosis of 50%, right coronary artery (RCA) mid 95% lesion, small left circumflex artery (LCX) and mildly reduced left ventricular function with inferobasal akinesis. The patient was referred for CABG.

Eight months after her CABG, she presented with unstable angina. Coronary angiography and a graft study revealed unchanged severe native vessel disease, patent saphenous vein graft (SVG) to RCA and patent SVG to first diagonal branch. Left internal mammary artery (LIMA) angiography revealed an 80% lesion at the LIMA-LAD anastomosis. Intensification of medical therapy was recommended at this stage.

The patient re-presented to the hospital for the third time 2 months later with acute coronary syndrome and high cardiac troponin. On this occasion, the coronary angiogram showed progression of her native disease over this short period. Her RCA is now 100% occluded at mid-segment ( figure 1 ). There was a new 70% narrowing in posterior descending artery branch (PDA) of the RCA, a new tight ostial LCX stenosis and LAD stenosis of 70%. At this point, a decision was made to perform a PCI to the RCA and its PDA branch with a drug eluting stent placed in each vessel. The patient continued to suffer from ongoing angina at rest and was brought for further intervention. This time, a successful PCI was performed for her LAD. She tolerated the procedure very well.

Right coronary artery total occlusion.

One month later, she re-presented with another NSTEMI. An angiogram showed patent SVG graft to PDA, but 90% in-stent restenosis of the PDA branch stent, a new 100% occlusion of the native RCA proximally ( figure 2 ) and critical LIMA-LAD anastomosis site stenosis ( figure 3 ). A balloon angioplasty was successfully performed for the PDA in-stent restenosis ( figure 4 ) and a stent was placed in the LIMA-LAD stenosis.

Posterior descending artery critical stenosis.

Left internal mammary artery to left anterior descending anastomosis site stenosis.

Posterior descending artery stent after baloon angioplasty.

Again, the patient presented with NSTEMI in 6 months’ time and angiogram showed a new 100% LMS lesion at the origin, RCA blocked, good SVG to RV branch with a new 90% in-stent restenosis within the PDA stent, LIMA to LAD occluded and SVG to diagonal was patent. EF at this time was 28%. A decision was made to continue medical treatment which included aspirin, clopidogrel, statin, β-blocker and ACE inhibitor.

Investigations

As this was a very unusual course of events with very rapidly progressing coronary lesions, the question of an underlying systemic cause for her CAD was raised. A homocystine level was negative and an autoantibody screening showed negative antinuclear antibody, extractable nuclear antigen, antineutrophil cytoplasmic antibody, antidouble-stranded DNA, and a positive β 2 glycoprotein-I. A repeat autoantibody screen again confirmed a high titre of anti-β 2 glycoprotein-I and anticardiolipin antibody. This was repeated after 12 weeks with similar results.

The patient was seen in the coagulation disorders clinic and a diagnosis of antiphospholipid syndrome was made based on the immunological profile, previous miscarriage and evidence of arterial thrombosis. She was then started on long-term warfarin and was to remain on aspirin for life, and clopidogrel was stopped after 6 months.

This is a very unusual report of APS diagnosed in a 61-year-old woman with recurrent acute coronary syndrome. It is unique because of the age of the patient and the nature of thrombotic events. APS usually presents with recurrent venous and arterial thrombosis. First presentation with CAD is extremely rare and only a few cases have been reported. 6

In patients with systemic lupus erythematosus (SLE), the presence of secondary APS is an independent factor for increased cardiovascular risks. 7 A Swedish cohort of 182 patients with SLE had also shown a very high prevalence of CAD among those who had APS. In that study, a positive antiphospholipid was independently related to the time to the first cardiovascular event. 8 This is probably a reflection of increased endothelial cell activation and damage. This case is more unique because of the extent and speed of the disease progression.

This endothelial dysfunction was assessed at the molecular level in patients with APS. Pro-inflammatory cytokines IL2, IL6 were found to be significantly higher in patients with APS coronary artery disease and this correlated directly with markers of endothelial dysfunction. 9

The association between the presence of antiphospholipid antibodies and the risk of cardiovascular events was studied in 2001 within the Honolulu heart programme. Anti-β 2 glycoprotein-I was associated with increased risk of stroke and myocardial infarction. 10 In another cross-sectional study, sequential patients admitted with suspected coronary artery disease were tested for the presence of APS antibodies. Patients with angiographically documented CAD had higher incidence of positive antibodies and those with positive antibody profile had more severe disease. 11

Learning points

• Consider antiphospholipid syndrome (APS) as a possible cause of aggressive coronary artery disease.
• Further research is needed to fully understand the extent of association of APS and proinflammatory mediators and atherosclerosis.
• More studies are needed to elaborate on any causal relationship between β 2 glycoprotein-I antibodies and coronary artery disease.

Contributors: ASA wrote the original script and edited the final manuscript. HY and TJK and CD participated in literature review and writing the manuscript. CD participated in literature review and writing the manuscript and is the principal physician in patient care.

Competing interests: None.

Patient consent: Obtained.

Provenance and peer review: Not commissioned; externally peer reviewed.