Coronary endothelial dysfunction

Article written and reviewed by Dr Luca Moretti

Coronary endothelial dysfunction is a pathophysiological phenotype in which the vascular lining does not adequately regulate tone, flow, inflammation, and thrombosis. It may involve the epicardial arteries, the microcirculation, or both, and may be present before a plaque becomes visible on angiography; it does not automatically constitute a cause of chest pain. It may be asymptomatic or contribute to ANOCA and INOCA. Diagnosis requires a defined functional test, not a clinical impression.
The invasive reference method is the response to intracoronary acetylcholine administered slowly in progressively increasing low doses, while vessel diameter and flow are measured. In intact endothelium, muscarinic activation promotes nitric oxide release and increased flow; if the endothelium is impaired, direct smooth-muscle contraction may predominate. The epicardial response and the microvascular response must be reported separately. The dose, infusion rate, and site of administration change the meaning of the test. A protocol that is not described cannot be interpreted rigorously.

Endothelial dysfunction is not synonymous with vasospastic angina. The latter requires an episodic syndrome and, during provocation, epicardial constriction greater than 90 percent associated with symptoms and ECG ischemia. Endothelial testing instead uses a submaximal stimulus to assess endothelium-mediated vasodilation. High doses administered rapidly may directly contract smooth muscle. A test positive for spasm does not selectively quantify endothelial function. The two phenotypes may nevertheless coexist.
The finding is valuable as an indicator of early vascular biology and as a possible component of non-obstructive ischemia, but must not be turned into an absolute prognostic label. Integrated assessment includes plaque anatomy, flow reserve, microvascular resistance, symptoms, and comorbidities; correcting risk factors is essential even if the test is not repeated. No drug is approved solely to normalize the acetylcholine response. Therapy is guided by overall risk and concomitant endotypes.

Etiology, risk factors, pathogenesis, and pathophysiology

The endothelium responds to shear stress, circulating mediators, and local signals by releasing nitric oxide, prostacyclin, and hyperpolarizing factors; these mediators dilate the vessel and inhibit platelet aggregation, leukocyte adhesion, and smooth-muscle proliferation. Endothelial homeostasis therefore extends far beyond regulation of diameter alone. The resulting tone reflects a dynamic balance with endothelin and other vasoconstrictors. Pulsatile flow and pressure continually modify the response. Clinical measurement captures only part of this biology.
Nitric oxide is produced from L-arginine by endothelial nitric oxide synthase and diffuses into smooth muscle, where it increases cyclic GMP and relaxation. Oxidative stress may inactivate it or uncouple eNOS, promoting superoxide production instead of NO. NO bioavailability therefore depends on synthesis, degradation, and the redox environment. An isolated plasma concentration does not measure this coronary balance. Precursor supplements do not necessarily restore function. Vasomotor testing assesses an integrated in vivo response.

ApoB-containing lipoproteins, smoking, hyperglycemia, and elevated blood pressure generate mechanical and oxidative stress, activate NF-kB, and increase adhesion molecules. Monocytes and platelets interact more readily with a dysfunctional surface; this proinflammatory phenotype links systemic exposures to the initiation and progression of atherosclerosis. Dysfunction may precede visible stenoses. Not every patient progresses at the same rate. The duration and intensity of exposure matter more than an occasional measurement.
Insulin resistance reduces PI3K-Akt-eNOS signaling while mitogenic and vasoconstrictor pathways may be preserved. Diabetes, visceral obesity, and elevated triglycerides promote inflammation, glycation, and glycocalyx disruption. Metabolic dysfunction simultaneously affects the epicardial and microvascular endothelium; glycemic control is not the only target. Weight, blood pressure, lipids, and physical activity act on the same substrate. Drugs are selected for demonstrated clinical benefits, not for an isolated biomarker.

Chronic kidney disease causes exposure to uremic toxins, phosphate, anemia, inflammation, and oxidative stress; autoimmune diseases add immune activation and potentially vascular effects of therapy. Aging reduces regenerative capacity and the response to shear stress. Endothelial reserve is therefore influenced by numerous noncoronary conditions. Diagnosis must place kidney function and inflammation in context. A treatable systemic cause must not be overlooked. The coronary finding does not justify simplistic causal attributions.
Sex hormones, pregnancy, menopause, and differences in microvascular structure modulate function, but the relationships are complex and nonlinear. Women are overrepresented in ANOCA cohorts, although endothelial dysfunction is not exclusive to females. Biological sex modifies probability and response together with age and risk factors. Hormone therapy is not a standard treatment for the coronary phenotype. Thrombotic risk must be assessed separately. Stereotypes must not replace objective measurement.

Acetylcholine stimulates muscarinic receptors on the endothelium and smooth muscle. With competent endothelium, vasodilation and increased flow predominate; if endothelial release is insufficient, the direct contractile effect becomes evident. This biphasic response explains why dose and infusion rate are essential: a slow low dose prioritizes the endothelial question. A rapid high dose tests susceptibility to spasm. Confusing the protocols produces conceptually incorrect diagnoses and treatments.

Pathological anatomy, classification, and progression

In the epicardial compartment, the change in diameter of angiographically smooth segments or those with mild plaque is observed during acetylcholine administration. A healthy coronary artery tends to dilate as flow increases; a dysfunctional response shows failure to dilate or constriction. Epicardial dysfunction is defined using thresholds that are not uniform across studies. Some cohorts consider any reduction from baseline. Others require constriction greater than 20 percent. The report must state the criterion used.
In the microcirculation, arteriolar diameter cannot be seen directly on angiography, so the response is inferred from the increase in flow measured by Doppler. An increase of less than 50 percent after low-dose acetylcholine is a widely used definition of endothelium-dependent microvascular dysfunction; the value depends on velocity and the diameter of the epicardial vessel. Measurement error is amplified in the flow calculation. A single threshold does not eliminate biological uncertainty.

The epicardium and microcirculation may be discordant. In a contemporary ANOCA cohort, classification based only on diameter would have missed some cases identified by flow, while using only flow would have missed other patients. Combined assessment therefore describes a more complete phenotype; this evidence comes from selected populations referred for testing. It does not estimate prevalence in the general population. The clinical relevance of mild abnormalities remains context-dependent.
Endothelial dysfunction, microvascular spasm, and adenosine-mediated microvascular dysfunction are not synonymous. Microvascular spasm requires symptoms and ECG ischemia during acetylcholine administration without epicardial spasm; endothelium-independent CMD is suggested by reduced CFR or elevated resistance during adenosine administration. Functional taxonomy separates the stimulus, compartment, and result. The same patient may have multiple endotypes. Each abnormality may require different therapy. Generic microcirculatory labels lose this information.

Dysfunctional endothelium promotes permeability, lipoprotein entry, leukocyte adhesion, and a prothrombotic state, creating an environment favorable to plaque formation. Intracoronary imaging may show early atherosclerosis not evident on luminography. Subclinical atherosclerosis and dysfunction share causes and reinforce each other. It is not always possible to establish which appeared first. Absence of stenosis does not mean a normal vessel wall. Prevention targets the common causal exposures.
Longitudinal studies have associated an abnormal endothelial response with anatomical progression and more cardiovascular events; these observations do not demonstrate that the test is a validated surrogate treatment target. Vascular progression also depends on LDL, blood pressure, diabetes, kidney function, and baseline plaque; a response may improve without eliminating accumulated risk. It may worsen without making an event inevitable. Individual prognosis must integrate many variables.

Endothelial function is dynamic and may change with smoking, exercise, diet, medications, and inflammatory status; this partial reversibility distinguishes the phenotype from a fixed anatomical lesion, but does not guarantee complete normalization. Temporal variability limits the use of repeated tests as a routine objective; repeating an invasive procedure requires a concrete clinical decision. Follow-up is based primarily on risk, symptoms, and function. A change in the clinical picture reopens the differential diagnosis.

Clinical manifestations

Endothelial dysfunction may be entirely asymptomatic and detected in physiological studies or during assessment of early atherosclerosis. It has no pathognomonic symptom. Clinical silence is consistent with its nature as a functional marker, but does not justify invasive population screening; testing is reserved for people in whom the result changes diagnosis or treatment. Overall risk is assessed with less invasive clinical tools. An incidental finding should be interpreted without alarmism.
When it contributes to ANOCA or INOCA, it may be associated with exertional angina, dyspnea, fatigue, or prolonged chest discomfort. Inadequate vasodilation limits the increase in flow required during activity, especially if elevated demand or structural CMD coexist. The exertional phenotype nevertheless remains nonspecific; diffuse stenosis, anemia, and heart disease produce similar symptoms. Ischemia testing must be technically valid. The mechanism cannot be inferred from the description of the pain.

Some patients report pain at rest or variable pain, particularly when vasomotor hyperreactivity and concomitant spasm are present. Isolated endothelial dysfunction must not, however, be used as a synonym for a vasospastic episode. Episodic pain with ST-segment changes requires specific testing and criteria for the epicardium or microcirculation. Low-dose testing assesses endothelial regulation. High-dose testing may reveal a separate spastic endotype. The report must preserve both conclusions.
Symptoms may persist after coronary angiography described as normal, leading to repeated assessments. Explaining that luminography does not measure function may correct false reassurance, but a functional diagnosis must not become a catch-all for every pain. Clinical-physiological consistency requires plausible reproduction, ischemia when present, and exclusion of alternatives. Reflux, chest-wall disorders, and pulmonary disease may coexist. Anxiety may amplify the burden without invalidating vascular disease. Management must address every documented component.

Migraine, Raynaud phenomenon, and inflammatory diseases are more frequent in some groups with vasomotor disorders, but are not diagnostic criteria for coronary endothelial dysfunction. A systemic association may guide history-taking and specialist collaboration; it does not allow a peripheral response to be automatically extrapolated to the coronary arteries. Vasoactive drugs taken for other conditions must be recorded. Overall benefit guides any treatment changes.
Reduced exercise tolerance may result from an inadequate coronary response, deconditioning, obesity, anemia, diastolic dysfunction, or pulmonary disease. Cardiopulmonary exercise testing helps quantify the limitation but does not selectively identify the endothelium. Multifactorial dyspnea calls for echocardiography, hemoglobin measurement, and respiratory assessment as appropriate. Attributing everything to the coronary test may delay useful treatment. Response to therapy does not retrospectively confirm the mechanism. Functional objectives must be measurable.

New or prolonged pain, or pain associated with diaphoresis, syncope, and instability, must be assessed as a possible acute event, irrespective of a previous endothelial diagnosis. Urgent reassessment includes ECG and troponin according to risk; endothelial dysfunction may coexist with plaque, thrombosis, or dissection. An old coronary angiogram does not exclude new disease. The usual pattern is useful only if critically compared with the current episode.

Investigations and diagnosis

Before invasive testing, symptoms, probability of epicardial disease, ventricular function, and the quality of previous tests are defined. CT coronary angiography or invasive angiography describes plaque and stenosis; FFR or nonhyperemic indices assess intermediate lesions. Endothelial function is studied after excluding an obstruction capable of explaining the clinical picture. Significant stenosis alters reference pressure and flow. The protocol must control medications and caffeine according to local procedures. Blood pressure and heart rate are stabilized before measurements.
One approach uses selective acetylcholine in the left anterior descending artery at doses equivalent to approximately 0.5, 5, and 50 micrograms, infused slowly; the ENCORE protocol used 2.16, 21.6, and 108 micrograms in the left coronary artery, with each dose infused over three minutes. These endothelial doses are not interchangeable without considering the vessel, concentration, and duration. Angiography and Doppler measurements are acquired at baseline and after each dose. The cumulative dose must be explicit. A final nitrate assesses maximal diameter and endothelium-independent function.

Epicardial diameter is measured with quantitative angiography in the same segments and projections. A normal response is dilation or at least the absence of significant constriction; an abnormal response ranges from failure to dilate to paradoxical constriction. The epicardial threshold is not universally standardized; many historical studies require a reduction greater than 20 percent. Recent cohorts classify any reduction from baseline. Comparisons among studies must respect this difference.
Coronary flow is estimated by combining Doppler average peak velocity and the area of the epicardial vessel. An increase of less than 50 percent after acetylcholine is the most commonly used criterion for the microvascular component. The flow response is strongly affected by the accuracy of the diameter, which is squared in the calculation; a stable wire position reduces variability. Motion, drift, and a weak Doppler signal must be reported. A technically inadequate value must not be forced into a category.

High-dose acetylcholine administered rapidly, for example up to 100 micrograms in the left coronary artery or 50 in the right in Japanese protocols, is intended to provoke spasm. Spasm provocation requires symptoms, ECG ischemia, and constriction greater than 90 percent for the epicardial phenotype; this response includes direct action on smooth muscle. On its own, it does not demonstrate selective endothelial dysfunction. A comprehensive procedure may perform both phases while reporting them separately.
Adenosine induces predominantly endothelium-independent hyperemia and allows measurement of CFR and resistance indices such as IMR. Low CFR or elevated IMR identifies a different microvascular component, often structural or endothelium-independent. Adenosine flow reserve complements but does not replace acetylcholine; reduced CFR may also result from high resting flow. Epicardial physiology must be excluded. Interpreting every value as an endothelial measure leads to therapeutic errors.

Brachial flow-mediated dilation, digital tonometry, and circulating biomarkers assess aspects of peripheral or systemic vascular function. Group-level associations with the coronary response exist, but individual concordance is insufficient for these methods to be considered equivalent. A peripheral test neither confirms nor excludes coronary dysfunction. PET and quantitative CMR measure perfusion and global reserve. They may document CMD but do not isolate the endothelial pathway. Clinical use must respect what each method actually measures.
Invasive acetylcholine testing is generally safe at experienced centers, but may provoke bradycardia, block, hypotension, severe spasm, and arrhythmias. Procedural control requires continuous monitoring and immediate availability of nitrate, atropine, pacing, and resuscitation; severe bronchospasm and major impulse formation or conduction disorders require specific caution. High-risk acute conditions, unstable heart failure, and severe dysfunction are stabilized first. Consent must explain which decision depends on the result.

The final diagnosis states the compartment, dose, threshold, and presence of associated endotypes. A proper report may describe epicardial dysfunction with preserved flow, microvascular dysfunction with preserved diameter, combined dysfunction, or a normal response. Explicit phenotyping prevents endothelial dysfunction, spasm, and CMD from becoming synonyms. Borderline results require caution and technical review. The differential diagnosis includes diffuse plaque, cardiomyopathies, and noncardiac causes. A single abnormality does not prove that every symptom results from it.

Treatment and prognosis

No therapy is approved solely to normalize the acetylcholine test. Treatment reduces exposures that damage the endothelium and addresses symptoms, atherosclerosis, and coexisting endotypes. The therapeutic strategy begins with smoking cessation, regular physical activity, blood-pressure control, a cardioprotective diet, and weight management; these measures have clinical benefits beyond the endothelial marker. Goals must be realistic and measurable. Rehabilitation supports adherence and functional recovery.
Statins reduce atherogenic lipoproteins, stabilize plaque, and have improved vasoreactivity in some small studies. Prescribing must nevertheless follow cardiovascular risk, LDL, and the presence of atherosclerosis, not a purported universal indication based on the test. The prognostic benefit derives primarily from reducing events in appropriate populations. A change in diameter is not a sufficient surrogate. Intolerance is managed with validated doses and alternatives. Aspirin is not automatically indicated in the absence of plaque or a prior event.

ACE inhibitors and angiotensin receptor blockers improve blood pressure, vascular signaling, and remodeling in appropriate settings. The 2024 ESC guidelines state that an ACE inhibitor may be considered for symptom control in patients with endothelial dysfunction, class IIa and level B. The ESC recommendation is not equivalent to evidence of event reduction specific to this phenotype. Kidney function and potassium must be monitored. Cough or angioedema directs the choice of alternatives. Normotension and hypotension limit titration.
In the WISE trial, quinapril improved microvascular reserve and angina in selected women with ischemia, non-obstructive coronary arteries, and microvascular dysfunction. Older studies such as TREND showed improvement in the vasomotor response in coronary artery disease; these physiological signals support plausibility but do not define a single treatment for everyone. Populations, endpoints, and protocols were heterogeneous. Individual benefit should be assessed through blood pressure and symptoms. Therapy is not prescribed merely to pursue a number.

The large WARRIOR trial, published in 2026, compared an intensive strategy using a high-intensity statin, an ACE inhibitor or angiotensin receptor blocker, and aspirin with usual care in 2,476 women with suspected ANOCA/INOCA. It did not reduce the primary clinical outcome, with a hazard ratio of 1.13 and p = 0.20. The neutral result warrants caution because of statistical power and crossover, and the trial was not selected for endothelial dysfunction. It does not justify automatic aspirin therapy. It does not negate individual indications for a statin or renin-angiotensin system blockade. It demonstrates the limitations of non-phenotyped empirical treatment packages.
Antianginal therapy depends on the associated endotype. Calcium channel blockers are indicated if spasm is documented; beta-blockers may be appropriate in CMD with elevated demand or resistance and no vasospasm; ranolazine is considered for persistent symptoms. Endotype-specific selection is more rational than an identical sequence for everyone. CorMicA improved angina and quality of life with stratified therapy. It did not demonstrate a reduction in major events. Tolerability and comorbidities guide every combination.

Antioxidant supplements, L-arginine, vitamins, and products described as nitric oxide stimulators lack sufficient evidence for routine use. Pharmacological doses may interact with medications or produce adverse effects. Outcomes-based medicine prioritizes smoking cessation, exercise, and control of LDL, blood pressure, and diabetes. Improvement in biomarkers does not guarantee fewer events. Over-the-counter products must be disclosed to the physician. Nitrates must not be combined with PDE5 inhibitors.
Follow-up means reassessing symptoms, functional capacity, adherence, and risk factors, not periodically repeating acetylcholine testing. New pain, troponin elevation, or functional decline calls for renewed anatomical and differential assessment according to the clinical picture. Clinical surveillance separates atherosclerotic risk from symptom burden; a highly symptomatic patient may have moderate risk. An asymptomatic patient with plaque and diabetes may be at high risk. Communicating both dimensions avoids inappropriate reassurance or alarm.

Prospective cohorts have associated endothelium-dependent vasoconstriction with plaque progression and more events. Prognostic strength varies according to the definition, population, and covariates. Individual prognosis integrates atherosclerotic burden, CFR, ventricular function, kidney function, diabetes, and inflammation; no single percentage applies to every positive test. An abnormal result identifies an opportunity for prevention. It does not inevitably predict myocardial infarction or death.

Complications

Endothelial dysfunction promotes the entry and retention of lipoproteins, leukocyte adhesion, permeability, and thrombotic imbalance, contributing to atherosclerotic progression; this biological complication often occurs without symptoms and may precede evident stenosis. The process nonetheless depends on cumulative exposure to apoB-containing particles and other factors. Dysfunction does not replace measurement of plaque. Lipid and blood-pressure prevention follows overall risk.
An inadequate microvascular response may limit the increase in flow and contribute to exertional ischemia, angina, or dyspnea, especially when elevated resistance and heart disease coexist. The perfusion limitation is not inferred automatically from endothelial testing; consistent symptoms or evidence of ischemia are needed when the question is INOCA. Global perfusion abnormalities may escape relative techniques. PET or quantitative CMR assesses complementary aspects.

Persistent angina and repeated procedures represent important clinical harm. Coronary angiography described as normal may lead to discharge without an explanation, while a generic diagnosis may lead to unnecessary polypharmacy. Diagnostic precision reduces both extremes; the patient must understand what was measured and what remains uncertain. Therapeutic goals include function and quality of life. Extracardiac causes are treated when documented.
Events such as myocardial infarction, stroke, heart failure, and death are more frequent in some cohorts with abnormal vasoreactivity, but arise from interaction with atherosclerosis and systemic risk. The observational association does not prove that the endothelium alone causes every event; confounding and patient selection affect estimates. Preventive therapies are chosen on the basis of validated clinical outcomes. Normalization of the test is not yet a standard therapeutic endpoint.

Ventricular arrhythmias and cardiac arrest are not defining complications of isolated endothelial dysfunction. If they occur during pain, epicardial spasm, severe ischemia, scar, channelopathy, or another cause must be sought. Assessment for vasospasm requires high-dose provocation with specific criteria; attributing an arrhythmia to mildly abnormal endothelial testing is dangerous. Monitoring and imaging are selected according to the case. An overlapping phenotype markedly changes prognosis.
Microvascular dysfunction and low reserve are associated with diastolic dysfunction and risk of HFpEF, but are not equivalent to the endothelium-dependent response. Elevated filling pressures may in turn reduce the perfusion gradient. The bidirectional relationship requires assessment of structure, function, and congestion. Dyspnea is not attributed to the endothelium by exclusion. Echocardiography and cardiopulmonary exercise testing clarify the phenotype. Heart failure is treated according to dedicated guidelines.

The diagnostic procedure may cause spasm, bradycardia, block, arrhythmia, or vascular complications, although the risk is low at experienced centers. Inappropriate selection exposes patients to risk without benefit. Procedural complications are prevented with protocols and immediate availability of nitrate, pacing, and resuscitation; the test is not a population screening tool. Severe bronchospasm and clinical instability require particular caution. Technical quality is an integral part of safety.
A false-positive result may turn physiological variation or an artifact into chronic disease; a false-negative result may leave a genuine mechanism unexplained. Different thresholds make it difficult to compare reports from different laboratories. Methodological standardization must specify the vessel, dose, duration, diameter, flow, and final nitrate. Borderline values should be communicated as such. Review of the signals may prevent repeated procedures. Clinical decisions must not depend on a number without context.

Empirical polypharmacy may cause hypotension, bradycardia, renal abnormalities, hyperkalemia, myalgia, or bleeding without proportionate benefit. Aspirin is especially problematic if prescribed solely for endothelial dysfunction in the absence of an atherothrombotic indication. Preventive iatrogenic harm is reduced by linking each drug to a validated objective. Laboratory tests verify safety and adherence. Adverse effects are distinguished from clinical progression. Simplifying unnecessary therapy may improve quality of life.