Endothelial Function Testing: Ready for Prime Time?

Abstracts & Commentary

Synopsis: Improvement in endothelial vasomotor function responses to a given therapy may or may not predict the results of large, randomized trials.

Sources: Stein, JH, et al. J Am Coll Cardiol. 2001;38: 1806-1813; Maxwell AJ, et al. J Am Coll Cardiol. 2001;38:1850-1858; Pullin CH, et al. J Am Coll Cardiol. 2001;38:1799-1805; Shinozaki K, et al. J Am Coll Cardiol. 2001;38:1821-1828; Chan N, et al. J Am Coll Cardiol. 2001;38:1814-1820; Kuvin JT, et al. J Am Coll Cardiol. 2001;38:1843-1849; Silber HA, et al. J Am Coll Cardiol. 2001;38:1859-1865.

There has been a remarkable increase in the number of reports of endothelial function assessment in cardiology journals over the past several years. In addition to the familiar use of brachial artery diameter or forearm flow studies, which have mostly superseded coronary angiographic testing for practical reasons, there are an increasing number of reports of plasma levels of endothelium-derived substances. The latter include ICAM-1, VCAM-1, PAI-1, TPA antigen, CRP, macrophage colony stimulating factor, and a variety of other molecules. Furthermore, we are beginning to see reports about the prognostic implications of disordered baseline endothelial vasomotion or biochemical markers with respect to adverse coronary artery events in subsequent years. In the December issue of the Journal of the American College of Cardiology, there are 7 articles dealing with endothelial function. These are briefly described in this commentary and should provide the reader with a sense of the robust research activity in this rapidly evolving field. Nevertheless, 20 years after Furchgott first described endothelium-derived relaxing factor (EDRF) and the subsequent identification that nitric oxide (NO) is the molecule responsible for modulating many endothelial functions (resulting in the 1999 Nobel Prize awarded to Furchgott, Ignarro, and Murad), it is still not known if assessment of endothelial function outside the research laboratory is relevant to the clinician, or has prognostic significance. Methodologic issues abound, particularly relating to the various methods of assessing vasomotor function.

Statins and Antioxidants: A study assessing the effect on endothelial vasodilator function of statins and antioxidants in hyperlipidemic patients was reported by Stein and colleagues. This report focused on older individuals (> 70 years of age) with an LDL cholesterol averaging 180 mg/dL and a total cholesterol of 262 mg/dL. Surprisingly, although baseline endothelial vasodilation in the brachial artery was abnormal, statin- induced lipid lowering with or without vitamin C or vitamin E for 12 months did not improve the abnormal endothelial vasomotor function. Stein et al concluded that aging alone has a powerful effect on depressing NO- mediated vasodilation, and that "morphologic and physiologic changes in the vascular wall in aging populations may preclude the beneficial effects of therapies that have been shown to improve endothelial vasomotor function in younger individuals with abnormal coronary artery disease (CAD) risk factors."

Uric Acid: It has long been known that uric acid levels are higher in populations with vascular disease. Maxwell and associates demonstrate that uric acid levels are related to vascular NO activity, and they speculate that elevations of uric acid are associated with the severity of cardiovascular disease. Furthermore, uric acid levels were inversely related to the degree of endothelial flow mediated vasodilation. In this study, there was a strong correlation between uric acid levels and abnormal flow mediated vasodilation with cardiovascular disease. Furthermore, pooled data from 6 clinical trials using L-arginine in a "medical food" (the Heart Bar) indicated that urate levels decreased and flow-mediated dilation increased with ingestion of this arginine supplement in subjects with coronary and peripheral artery disease or hypercholesterolemia who ingested the supplement for 1-2 weeks. In addition to the demonstrated relationship of serum uric acid levels to vascular NO activity and endothelial vasomotor function, there was a correlation between serum uric acid and cardiovascular risk factors. Furthermore, Maxwell et al hypothesize that with vascular disease or CAD risk factors, oxidative stress is increased and NO activity is decreased, resulting in increased xanthine oxidase activity, which removes the brake on uric acid production.

Homocystine and Folate: In spite of the recent report indicating that folic acid and B vitamins decrease restenosis,1 a report from Pullin and associates indicates that while homocystine (HC) levels are inversely related to plasma folate concentrations and were reduced by dietary or supplemental folate, there was no improvement in endothelial function in a large group of healthy subjects. In addition, genotype assessment indicated the expected variance of individuals with MTHFR gene mutations; these individuals with higher HC levels had a similar beneficial response to dietary or supplemental folate, and the same lack of improvement in brachial artery vasodilation after 4 months of therapy with either folic acid supplements or increased dietary folate.

Insulin Resistance: An elegant Japanese study from Shinozaki and associates of normal subjects stratified into tertiles of insulin sensitivity demonstrated that insulin-resistant subjects had abnormal vascular oxidative stress and pteridine metabolism, which correlated with abnormal coronary endothelial function.

Cardiovascular Risk Factors and Their Correlation with Reactivity: A British study of a large cohort of healthy individuals assessed a wide variety of CAD risk factors, as well as Framingham risk category. Lipoprotein particle size and standard lipid parameters were measured. Forearm acetylcholine and bradykinin responses were assessed; L-NMMA, an antagonist of arginine, was given to measure basal NO production. The major correlates of abnormal vascular reactivity were hyperlipidemia, high body mass index, and smoking. Furthermore, Framingham risk status was predictive of stimulated endothelial vasodilation, with lesser responses in higher-risk subjects. A great deal of variation in the vascular response to various agents was noted and was generally unexplained. Chan and colleagues concluded that lipids and smoking, but not blood pressure or lipoprotein particle size, are associated with perturbations in the NO pathway in these middle-aged subjects. A high Framingham risk score was predictive of reduced basal as well as stimulated endothelial dependant vasodilation.

Nuclear Stress Testing and Endothelial Function: An interesting report by Kuvin and associates evaluated brachial artery diameter responses to forearm occlusion in low-to-medium CAD risk in subjects referred for myocardial perfusion stress imaging with sestamibi. Impaired brachial artery dilator responses were arbitrarily set at less than a 10% increase over baseline. An abnormal brachial artery response was predictive of an abnormal exercise perfusion test or multiple CAD risk factors with normal imaging. On the other hand, a normal response (brachial artery dilation > 10%) was predictive of a normal stress test (38 of 40); 21 of 23 subjects with a positive nuclear exercise test had flow-mediated dilation (FMD) < 10%. Kuvin et al suggest that while further investigations are indicated, brachial artery flow mediated dilation may be a useful screening test in low-risk individuals, in whom a normal response (> 10%) might make exercise testing unnecessary.

What is the Cause of Flow Mediated Dilation? A study by Silber and colleagues demonstrated that vascular wall shear stress is the primary determinant of arterial flow mediated vasodilation induced by peak hyperemia in normal subjects using phase-contrast magnetic resonance angiography. Endothelial vasodilator responses were linearly related to the stimulus. Smaller arteries had proportionally more dilation than larger arteries. The study was previously reported at the 2000 American College of Cardiology Scientific Symposium. This is an important observation because it demonstrates that flow mediated vasodilator responses are linearly proportional to systolic wall shear stress, which appear to be greater in small arteries. Silber et al suggest that this technique may further improve our ability to noninvasively use endothelial vasomotor testing to differentiate normal from abnormal responses.

Comment by Jonathan Abrams, MD

In 1991, I wrote an editorial entitled, "Should Cardiologists be Endotheliologists?"—that question is no longer arguable. It is imperative that clinicians who deal with cardiovascular disease and its prevention become familiar with some of the large amount of information that has accumulated regarding endothelial function. The fact that a single issue of a major cardiology journal should have so many reports dealing with endothelial vasomotor function attests to the importance of this subject. However, there are several cautionary notes:

• Improvement in endothelial vasomotor function responses to a given therapy may or may not predict the results of large, randomized trials. Thus, one finds a positive correlation between endothelial function assessment with lipid lowering and the use of angiotensin-converting enzyme inhibitors and important trials using these compounds (4S, CARE, LIPID, etc, as well as HOPE). On the other hand, there are many positive reports using endothelial testing with estrogen or hormone replacement therapy in postmenopausal women, and even more studies using vitamin C and or vitamin E in a variety of conditions; in the aggregate, these demonstrate improvement in vasodilator responses. However, the existing clinical trial database with estrogen or antioxidant vitamins has been disappointing, and today one cannot advocate either adjunctive therapy for patients at risk for or having vascular disease. Furthermore, the report by Stein et al indicates that other factors, such as advanced age and other CAD risk factors, are variables that may importantly modulate the endothelial responses to a specific intervention, such as statin therapy.

• What is the gold standard for endothelial testing? A variety of techniques are used, most commonly forearm occlusion with assessment of brachial artery diameter by ultrasound. Some studies rely on forearm blood flow or resistance measurements. In addition to the brachial artery, the coronary arteries have been frequently assessed with acetylcholine or, less often, bradykinin stimulation, but these invasive tests cannot be practically repeated. Also, there is no standardization of equipment, methodology, or consensus as to the level of vasodilation that is "normal" or abnormal.

Does it make a difference? The few prospective studies that have been reported to date, correlating baseline endothelial vasomotor responses with subsequent clinical events, are positive and suggest that such an approach is predictive and can be used in clinical medicine. Obviously, far more experience is necessary involving many more subjects, with widespread agreement upon the appropriate technique and responses for endothelial response or serum markers.

An increasing number of reports, none cited in this discussion, are emerging assaying a variety of endothelial derived compounds, such as plasminogen activator inhibitor (PAI-1), TPA, adhesion molecules, etc. We do not know whether measurements of abnormal endothelial derived substances correlate with abnormal vasomotor function. There may be marked discrepancies between such markers and simultaneous vasomotor functional testing, but there are no available data for these important comparisons. Standardization of assays and appropriate cut points for abnormal levels will be essential. Futhermore, it is clear that in acute coronary syndromes, endothelial activation can be readily documented with serum marker testing, but it is not so clear as to whether such markers are predictive of outcomes in chronic CAD.

In conclusion, it behooves the physician to be aware of what is going on in this area but to also keep a reserved and even skeptical attitude regarding results of provocative small studies that have not yet been supported by traditional randomized trials. Overall, this area of research has made major contributions to our understanding of the physiology of vascular diseases, as well as the preclinical state. Endothelial function testing may indeed be ready for prime time!


1. Schynder G, et al. N Engl J Med. 2001;345:1593-1600.

Dr. Abrams is Professor of Medicine, Division of Cardiology, University of Mexico, Albuquerque, NM.