OUP user menu

Current diagnostic concepts to detect coronary artery disease in women

Verena Stangl, Vanessa Witzel, Gert Baumann, Karl Stangl
DOI: http://dx.doi.org/10.1093/eurheartj/ehn047 707-717 First published online: 12 February 2008


The goal of non-invasive diagnostic testing is to detect coronary artery disease (CAD) timely and with optimal fidelity. In women, the accuracy of most functional diagnostic investigations is diminished compared with men. To minimize the sex-related bias in detection of CAD, there is a need for a non-invasive test strategy capable of risk-stratifying women, and of identifying subsets of patients in need of further invasive testing. The present review discusses the validity of classical diagnostic procedures for CAD in women, and addresses the relevance of newer techniques in the context of a female-specific diagnostic approach.

  • Women
  • Coronary artery disease
  • Diagnostic


Cardiovascular diseases are the main cause of death in industrialized countries, in women as well as in men. Coronary artery disease (CAD) represents 23% of all deaths in women.1 There is compelling evidence that women with CAD experience worse outcomes than men, irrespective of age.2,3 One explanation for this sex-specific difference is that women present with more advanced disease, owing to lack of early recognition and management. Accurate diagnostic assessment may represent the first step towards improving the outcome for women at risk. However, non-invasive detection of CAD in women is challenging due to a number of reasons, beginning with the fact that women with chest pain demonstrate a lower prevalence of CAD (Table 1).4 Their symptomatology, furthermore, is less predictive and more often atypical when compared with men.5,6 Women are older at the time of initial diagnosis, and age-related comorbidities limit exercise tolerance. As a result, diagnostic procedures are less accurate in women than in men, and consequently fail to detect coronary heart disease at an early time point: resulting in more serious consequences in follow-up.7

View this table:
Table 1

Factors affecting accuracy of diagnostic stress testing in women

Lower prevalence of coronary artery disease
Higher prevalence of non-ischaemic chest pain (microvascular abnormalities, mitral valve prolaps)
Less predictive symptomatology
Limited exercise tolerance due to older age at initial diagnosis
Different response to exercise than men
 Limited exercise capacity (mostly due to older age)
 Lower peak exercise values
 Lesser increase in left ventricular ejection fraction
 Increase in cardiac output by enhancing end-diastolic volume
 Inappropriate catecholamine release
Hormonal influences of oestrogens mimicking digitalis-like false positive ECG response
Anatomical differences affecting stress test results
 Female breast attenuation artefacts
 Smaller coronary artery size
 Smaller left ventricular chamber size
Higher prevalence of single vessel disease

The purpose of the present review is to evaluate the accuracy of diagnostic procedures for CAD in women and to refine a female-specific diagnostic approach.


Search for published data

We performed literature searches in English in the PuBMed Medline database (www.ncbi.nlm.nih.gov). Data on the significance of various cardiovascular techniques for the diagnosis of coronary artery disease in women were relevant for this review. We used the following medical subject heading terms: ‘women’ OR ‘gender’ AND ‘coronary artery disease’ AND ‘risk stratification’ (477 hits) OR ‘exercise ECG’ (1832 hits) OR ‘stress echocardiography’ (801 hits) OR ‘myocardial scintigraphy’ (2780 hits) OR ‘myocardial perfusion imaging’ (1997 hits) OR ‘calcium score’ (424 hits) OR ‘computed tomography angiography’ (1253 hits) OR ‘magnetic resonance imaging’ (7125 hits) OR ‘intima media thickness’ (447 hits) OR ‘endothelial function’ (1940 hits) OR ‘flow-mediated dilation’ (154 hits) OR ‘microvascular dysfunction’ (277 hits) OR ‘syndrome X’ (671 hits). After eliminating ineligible studies by review of titles and abstracts, we reviewed the full text of 384 articles.

We selected studies on at least one of the above-cited tests that have been performed only in women, or that presented results separately by gender. We included studies investigating the respective diagnostic technique in women (and in both women and men) with chest pain. We also reviewed meta-analyses on the subject, as well as the reference lists of review articles and eligible studies to complete search data. The searches covered a time period from inception to December 2007.

Risk stratification

Prevalence, clinical symptomatology, risk factors

The prevalence of CAD in women with chest pain is about 50%, compared with 80% in men.4 Half of the women with chest pain referred to coronary angiography in the Coronary Artery Surgery Study (CASS) had minimal or no coronary stenosis—in comparison to 17% in men.8 Overall, exclusion of CAD is observed five times more frequently in women than in men referred for coronary angiography.9 Reasons for this discrepancy are the higher prevalence in women of non-ischaemic chest pain, as well as less common causes of ischaemia such as vasospastic angina and cardiac syndrome X.5 Cardiac syndrome X characterized by exercise-induced angina and transient ischaemic ST-segment depression in the absence of epicardial obstruction is much more common in postmenopausal women than in men, and has been linked to oestrogen deficiency.10,11 Microvascular abnormalities caused by dysfunction of endothelial as well as vascular smooth muscle cells may be responsible for myocardial ischaemia in patients with syndrome X.10,12,13

There are important sex-related differences in clinical presentation. Whereas initial manifestations of CAD in men are more often myocardial infarction and sudden cardiac death, women are more likely to experience angina as their first symptom. However, chest pain or typical angina is less predictive for CAD in women than in men.5,6 Symptoms of angina in women frequently do not correspond to classical descriptions originally formulated for men. Women more often present with unusual fatigue, dyspnoea, epigastric discomfort, and back, jaw, or neck pain than do men.14 These more likely ‘atypical’ symptoms may be misinterpreted by both patient and doctor, leading to delayed diagnosis and intervention, and possibly contributing to less desirable outcomes. The new onset of symptoms and their relationship to physical activity have been suggested to possess more clinical relevance for women than does their localization.15 In women with suspected but yet undiagnosed CAD, evaluation of chest pain accordingly represents a critical step in the care.

There is some evidence that cardiovascular risk factors predict the existence and extent of CAD at a higher percentage in women than in men. Among the elderly, hypertension is not only more common in women than in men with CAD, but is also a stronger predictor of CAD in women than in men.5,16,17 Although smoking is still more prevalent among men, an increase in total lifetime intake of tobacco appears to entail a greater increase in risk for women than for men.5,18 With regard to lipid disorders, there is some evidence that low HDL cholesterol levels are a better predictor of the risk of CAD in women than is a high level of LDL.19,20 Triglycerides appear to influence the risk of CAD in women more than in men. The presence of diabetes mellitus is an explicitly more powerful predictor of CAD and its prognosis in women than in men.5,21,22 Diabetes and/or peripheral arterial disease are considered to be CAD risk equivalents, and women with these diseases are therefore classified as high-risk. The metabolic syndrome is a clustering of the risk factors known to have the ability to predict incident cardiovascular disease or diabetes.23 Research from the Nurses' Health Study (NHS) demonstrated that the prevalence of the metabolic syndrome was 41% in women with CAD, compared with 18% in disease-free controls.24 There is some evidence that the metabolic syndrome is a stronger predictor of CAD in women than in men.24 C-reactive protein appears to further improve cardiovascular risk classification in women, particularly in those with an intermediate 10-year risk.2527 Among women with or without the metabolic syndrome, C-reactive protein levels add clinically relevant prognostic information concerning future vascular risk.28 Screening for novel risk factors—such as C-reactive protein, as well as subclinical atherosclerosis—can further refine the assessment for cardiovascular risk in women.

Risk scores

In accordance with current guidelines regarding the algorithm of non-invasive testing for women, initial test choice should be guided by classifying women into symptomatic or asymptomatic and, further, into low, intermediate, or high pretest risk categories.29,30 Scoring systems may help to assign women to the respective risk group and to select those suitable for further diagnostic testing.31 The SCORE risk chart, which was developed using data from 12 European cohort studies, and which includes 205 178 subjects, separately estimates the 10-year risk of a first fatal atherosclerotic event in women and men. Very recently, the Reynolds Risk Score, a female-specific score derived from 25 000 participants of the Women's Health Study (WHS), has been suggested to improve the accuracy of cardiovascular risk prediction. Beyond classical risk factors, this model further includes high-sensitivity C-reactive protein and the family history of premature cardiovascular disease, both independent risk factors in women.32

Additional tools for risk stratification

Intima–media thickness

Carotid intima–media thickness (IMT), which can be easily and non-invasively assessed, has well-recognized value as a surrogate marker for the atherosclerotic burden, and has been positively linked to the presence and extent of CAD in both women and men.3335 IMT increases markedly with the number of risk factors, and may be considered to represent a comprehensive picture of the alterations caused by risk factors over time on the arterial wall.33

There is some evidence for sex differences in IMT.3436 Carotid IMT is smaller in women than in men,34,37 a phenomenon attributed to a slower increase of IMT with age in women. Sex differences disappear in subjects older than 65 years and in patients with CAD.34,38 Kablack-Ziembicka et al.34 found that IMT is associated with higher accuracy in predicting the likelihood of CAD (≥50% stenosis) in mostly symptomatic women. Although increased IMT (with a cutoff of 1.07 mm in women and 1.15 in men in this study) had high positive predictive value in women and men, specificity and sensitivity in women were much higher. The risk of CAD in women with IMT>1.07 was more than seven times that of women with IMT below this threshold.34

Although the significance of this method is not clearly defined in the guidelines, IMT assessment may be a good adjunct to clinical risk stratification and may help in decision-making for further diagnostic testing, particularly in asymptomatic women with intermediate cardiovascular risk.

Peripheral endothelial function

Endothelial dysfunction characterized by reduced production and/or bioavailability of nitric oxide (NO) is one of the earliest events in atherogenesis. Peripheral endothelial dysfunction assessed by depressed flow-mediated dilation (FMD) of the brachial artery has been suggested as risk stratification for women as well as for men.39 In patients with chest pain (40% of them women), forearm endothelial dysfunction was a sensitive indicator of CAD, but demonstrated poor specificity and an inability to predict the severity of CAD.40 With account taken of the caveats in FMD measurement—i.e. of the endogenous, exogenous, and environmental factors that may influence test results—assessment of endothelial-dependent vasodilation by ultrasound of the brachial artery evidently represents a sensitive method associated with an excellent negative predictive value for CAD.41 Although FMD has additionally been proposed as predictor of future cardiovascular events, recent data demonstrated that this methods adds little to the prognostic accuracy of traditional risk factors in adults (59% women).42,43

Exercise electrocardiography

Exercise ECG is the classical non-invasive initial investigation for diagnosis of stable CAD. An exercise tolerance test enables assessment of functional capacity and offers the advantages of wide availability and low initial costs.

Exercise ECG has long been recognized as being less sensitive and specific in women. The specificities with regard to epicardial obstruction ranged from 74 to 89% for men, and from 33 to 73% for women. This reflects the wide variability of published reports, primarily due as it is to differing clinical situations (i.e. symptomatic vs. asymptomatic, and high-risk vs. non-high-risk screening) and the resulting different pretest probabilities.4449 A comprehensive meta-analysis by Kwok et al.48 investigated the accuracy of stress tests for the diagnosis of CAD (epicardial obstruction of >50% to >70%, depending on the study) in symptomatic and asymptomatic women. A mean specificity of 70% was found for women (77% for men), which indicates that at least 30% of positive electrocardiograms in women are false positive with regard to epicardial obstruction.48 It has been suggested that oestrogen status and, especially, oestrogen replacement therapy may play a role in producing these false positive results. Using angiography as well as non-angiography-based methods in comparing the specificity of exercise ECG in 1880 men and 1818 women with suspected CAD (symptomatic and asymptomatic), Morise and Beto50 reported that premenopausal women had greater specificity than women on oestrogen replacement therapy, and that the latter demonstrated lower specificity than did postmenopausal women not undergoing oestrogen replacement.

Sensitivities are generally low in both genders: compared with epicardial obstruction, sensitivities of about 63–68% for men and 53–61% for women are consistent with the literature.48,51 The positive predictive value is in most studies below 50% for women, and about 70% for men.6,49

The reasons for lower accuracy of ECG stress testing in women are not well understood. It has been discussed that anatomical differences as well as autonomic and hormonal influences may play a role in sex-related discrepant findings.52 Oestrogen, with its molecular similarities to digitalis, may cause the digitalis-like false positive ECG response that has been shown to vary during the menstrual cycle and in response to oestrogen therapy.5355 In addition, women are more likely to experience short or inadequate duration of exercise stress, which results in lower sensitivity for the exercise treadmill. Other factors that may reduce accuracy of stress tests in women include the higher prevalence of microvascular disorders and/or vasospasm possibly associated with ST-segment depression without subsequent angiographic detection of epicardial obstruction.8,1012

There is some evidence that multivariate approaches, focusing not only on ST-segment depression in exercise ECG, but also considering additional parameters such as exercise capacity, maximum heart rate, heart rate recovery, ST/heart rate relation, ST recovery time, and QRS score, may improve the positive predictive value in women.46,47,56 Alexander et al.57 reported that the Duke Treadmill Score based on exercise time (or capacity), ST-segment deviation, and treadmill angina effectively stratifies diagnostic and prognostic risk in women. The Duke Treadmill Score served even better in women than in men for excluding CAD, with fewer low-risk women demonstrating significant CAD (≥1 vessel with ≥75% stenosis) (20 vs. 47% in men).57 In contrast, high-risk scores predicted significant CAD (≥1 vessel with≥75% stenosis) in 89% of women, and 98% of men. Although the Duke Treadmill Score may be methodologically criticized on the grounds that low-risk patients underwent only exercise tests and not cardiac catheterization, its positive predictive value may be improved by consideration of several parameters in women as well. Accordingly, the Scientific Statement of AHA, which considered the role of non-invasive testing in the clinical evaluation of women with suspected CAD, stressed that simple exercise scores integrating multiple parameters (e.g. the Duke Treadmill Score or ST/heart rate index) can improve the predictive value of exercise tests in women.30

In toto, exercise ECG is associated with a poor positive predictive value but a relatively better negative predictive value in women. A positive test result implicates that further diagnostic approaches are necessary. Although exercise ECG is the recommended initial diagnostic test for symptomatic women with an intermediate to high probability of CAD, women are less likely to undergo an exercise ECG than are men.58

Stress echocardiography

Stress echocardiography during dynamic exercise or pharmacological stress offers several advantages over exercise ECG: including superior diagnostic performance, the ability to localize areas of ischaemia, and the option of performing stress testing in patients unable to exercise adequately. Among the few published studies investigating the validity of stress echocardiography in women, most sparsely show sex-related differences in cardiac response to this test. Although sensitivities are reportedly slightly worse for women than for men, specificities appear to be comparable in women.59 Three studies performed in symptomatic women found specificities of exercise echocardiography for the diagnosis of epicardial obstruction (>50% stenosis at coronary angiography) ranging from 81–86% and sensitivities, from 80 to 88%.6062 There is some evidence that sub-maximal exercise (attainment of <80% of age-predicted maximal heart rate), single-vessel coronary artery disease, and moderate stenosis correlate with false negative exercise echocardiographic results in women.62

Pharmacological stress echocardiography with dobutamine, adenosine, or dipyridamole is particularly indicated in women with limited exercise capacity. A recent meta-analysis evaluated more than 80 studies with regard to accuracy of pharmacological stress testing for CAD. All patients included in this meta-analysis underwent coronary angiography. Of the only five studies that included sex-specific evaluations, overall sensitivity and specificity was 76 and 86%, respectively, for women and not significantly different from the corresponding values for men: 78 and 89%.63 A more recent study that prospectively compared the accuracy of dobutamine stress echocardiography with epicardial obstruction (>70% coronary stenosis) in a sex-related manner reported even higher sensitivities and specificities for women (90 and 85%) than for men (80 and 77%) at intermediate risk for CAD.64 Interestingly, it appears that this stress form is less well tolerated by women: a study investigating the impact of gender on hypotension induced by dobutamine stress echocardiography found higher incidence of hypotension (36%) in women than in men (24%).65

There is some evidence that adenosine and dipyridamole stress echocardiography results demonstrate lower sensitivity but higher specificity in women than does stress echocardiography performed with dobutamine.66 Using dipyridamole stress testing, Masini et al.67 reported a sensitivity of 79% and a specificity of 93% for diagnosis of significant CAD (>70% epicardial obstruction) in women with chest pain. Technical advantages of stress echocardiography over stress perfusion imaging in women include absence of breast-attenuation artefacts and the possibility of visualizing the mitral valve in women with chest pain.

Taken together, exercise and pharmacological stress echocardiography are apparently effective and accurate in detecting CAD in women, in a manner similar to that demonstrated for men. Accordingly, a recent consensus statement by the AHA contained a recommendation for exercise stress echocardiography for symptomatic women with intermediate to high pre-test probability of CAD, and for dobutamine stress echocardiography for women with normal ECG who were incapable of exercise.30

Radionuclide myocardial perfusion imaging

Myocardial perfusion imaging with exercise or pharmacological stress has been shown to add incremental value in risk stratification of women with intermediate clinical pre-test likelihood of CAD. It has been generally accepted that exercise thallium-201 scintigraphy is associated with lower sensitivity and specificity in women than in men. A comprehensive meta-analysis of exercise testing performed to detect significant epicardial obstruction (21 studies with a total of 4113 women) disclosed that the accuracy of thallium imaging is higher in men, with a sensitivity and specificity of 85 and 85%, respectively, compared with 78 and 64% in women.48 Smaller left ventricular chamber size in women appears to be one reason for the lower accuracy of this stress test in women. When exercise thallium results are adjusted for ventricular size, no differences in accuracy between men and women become apparent.68 Another reason for the less eloquent results includes enhanced attenuation of myocardial activity by the female breast, which particularly interferes with imaging abnormalities in the territory of the left anterior descending coronary artery.52,69 There is some evidence that technetium-99m sestamibi imaging is superior to thallium with regard to the problem of breast attenuation. Taillefer et al.70 accordingly reported sestamibi specificity of 82.4% for CAD diagnosis (>70% stenosis), compared with 58.8% for thallium-201 imaging in women. Yet another cause for diminished accuracy in women is that the higher prevalence of single-vessel CAD adversely affects the diagnostic accuracy of stress imaging.71 A study using thallium-201 imaging demonstrated sensitivity for single-vessel disease that was lower in women than in men.71

Perfusion imaging studies included in the meta-analysis of Kwok et al.48 incorporated data from older literature with older imaging techniques. With the use of gated single photon emission computed tomography (SPECT), simultaneously derived information on perfusion and function can assist in enhanced differentiation of attenuation artefacts from infarcts. Thus, two studies using gated SPECT imaging with technetium-99m sestamibi clearly improved the previously reported specificity for detection of CAD (>50% stenosis) in women (with a pretest probability of CAD of 50–75%) to 92 and 91%, respectively.7172

Pharmacological stress testing with adenosine and technetium-99m sestamibi SPECT myocardial perfusion imaging—which has become the most common stress test in both women and men who are unable to exercise—is apparently associated with high sensitivity and specificity in women, independently of presenting symptoms and pre-test likelihood of the disease.7375

Coronary calcium score

The coronary calcium score (CCS) assessed by cardiac computed tomography (CT) has been discussed as method for early detection of coronary atherosclerosis.76,77 Coronary artery calcium (CAC) is almost always present in cases of angiographically significant CAD, and thus demonstrates great sensitivity. Specificity, however, is low, since detection of calcification implies atherosclerosis, but not necessarily the presence of haemodynamically relevant stenosis. In the largest single study to date, with 1764 consecutive patients (including 539 women) being evaluated with coronary angiography for clinical reasons, the sensitivity of CCS was reported to be 99%, with equally high sensitivity for men and for women.78 Among the 41% women with normal coronary angiography, no CAC was detectable. Moderate or high CCS, however, was associated with greater prevalence of relevant CAD in women.78 A zero CCS score furnishes very strong evidence against CAD in patients with pre-test CAD likelihood of<50%. It must nevertheless be emphasized that low or zero CCS results do not exclude clinically important obstructive CAD. A recent study in 231 consecutive patients with chest pain (45% women) found obstructive CAD in one-quarter of patients with zero-to-low CCS.79 There is some evidence that the amount of CAC may be predictive of coronary events.80

On the whole, available data on coronary calcium and risk of CAD in women are sparse. Since high negative predictive values take centre stage, this approach may accordingly substantiate the diagnosis in selected patients with intermediate risk.

Computed tomography coronary angiography

CT coronary angiography using multi-slice spiral CT (MSCT) after intravenous contrast-agent application is a promising technique in terms of non-invasive imaging of coronary arteries. It includes the possibility of examining the arterial wall and plaque morphology. Rapid technical advances from 4- to 16- to 64-slice CT and beyond have led to significant improvement in the diagnostic accuracy of coronary MSCT.81 One of the key advantages of this method is its high negative predictive value: which—in recent, mostly small studies using 16- or 64-slice CT (total number of included women<400)—ranges from 95 to 100% among patients at intermediate risk of having CAD.8289 Sensitivities and specificities for detection of epicardial obstruction have been reported to be from 90 to 94% and 75 to 97%, respectively, for studies using 64-detector scanning.82,8691 In patients with high risk of CAD, however, the accuracy of MSCT is apparently still suboptimal. Recent guidelines for diagnosis and assessment of stable CAD recommend the use of non-invasive CT angiography for patients with low pre-test probability of disease, with non-conclusive exercise ECG or stress imaging tests.91 There is likewise some evidence that MSCT may provide an attractive tool for ruling out coronary aetiology in patients who present to the emergency unit with acute chest pain. Two prospective studies with a total number of 202 patients (98 women) reported that the absence of coronary plaque or significant stenosis on MSCT was associated with an excellent negative predictive value, up to 100%, for subsequent diagnosis of acute coronary syndromes.92,93

Two recent studies directly compared the diagnostic accuracy of 64-slice CT coronary angiography in women vs. men.88,89 Whereas Pundziute et al.,89 in investigating 103 consecutive patients (52 women) with known or suspected CAD, found no significant differences between women and men in sensitivities and specificities of MSCT angiography compared with invasive coronary angiography, a study by Meijboom et al.89 (n = 402 symptomatic patients, including 123 women) reported significantly lower diagnostic accuracy (88 vs. 96% in men), specificity (75 vs. 90% in men), and positive predictive value (81 vs. 95% in men) in women vs. men.88,89 One small study reported that female sex was one of the most important predictors of non-assessable segments (odds ratio 2.6), presumably due to smaller coronary-artery size.94 In the face of the small total number of women included in the studies, the relevance of CT coronary angiography for women requires further verification.

A notable restriction of MSCT results from its elevated radiation exposure, higher than that of conventional diagnostic coronary angiography: a factor that must particularly be taken into account in women.95 The effective dose may amount up to 15 mSv in women, which may give rise to associated breast-irradiation issues.96 Further limitations include reduced assessment of luminal narrowing in regions of calcified atheromatous plaques. Other constraints are imposed by metallic implants in the form of coronary stents, surgical clips, sternal wires, pacemakers, and prosthetic valves.97

In conclusion, MSCT coronary angiography may represent an attractive method of excluding CAD in selected women with low-to-intermediate probability of the disease, and for whom a false-positive stress test result is suspected. The newer generation of 256-slice CT with improved spatial and temporal resolution promises substantial future improvement of image quality that may allow more precise evaluation of coronary stenosis.

Cardiac magnetic resonance imaging

Cardiac magnetic resonance imaging (MRI) has recently emerged as a non-invasive imaging technique that provides information on cardiac morphology, myocardial function, viability, metabolism, as well as coronary morphology. This method is appealing because it offers high spatial resolution and does not expose patients to ionizing radiation. Myocardial perfusion can be estimated at rest by first-pass perfusion MRI with gadolinium, and during pharmacological vasodilation (adenosine) or dobutamine stress MRI with promising results.98103 To date, only a small number of studies have evaluated MRI for diagnosis of CAD in women. The Multi-Ethnic Study of Atherosclerosis assessed myocardial perfusion using MRI in relation to cardiac risk factors in asymptomatic women and men. Coronary vasoreactivity was reduced in asymptomatic patients with a greater coronary risk factor burden.104 In this study, both resting and hyperaemic myocardial blood flow (MBF) was lower in men than in women. The significantly higher MBF and perfusion reserve in female patients remained, even after menopause and hormone use were taken into consideration: which suggests that gender difference may not be completely explained by the direct effect of estrogens.104 The sensitivity and specificity values of perfusion MRI compared with coronary angiography for diagnosis of CAD range from 65 to 92% and from 76 to 100%, respectively.105

Dobutamine stress MRI appears to be well suited for overcoming the limitation of dobutamine stress echocardiography. Nagel et al.100 provided the first report on the efficacy of high-dose dobutamine and atropine stress MRI for detecting CAD (29% women included). In comparison to stress echocardiography, Nagel found higher diagnostic accuracy (sensitivity and specificity of 86%) for stress MRI. There is also some evidence for the utility of stress MRI in establishing cardiac prognosis. Hundley et al.,106 after investigating 153 patients (44% women), accordingly reported that patients with negative dobutamine stress MRI had a cardiovascular-occurrence, free-survival rate of 97%.

The concept of myocardial late enhancement in contrast-enhanced cardiac MRI has recently been established for the assessment of myocardial viability.107 A prospective trial including 47 patients (9 women) revealed that stress first-pass perfusion MR imaging followed by delayed enhancement is an accurate method for evaluation of significant coronary stenosis in patients with suspected or known CAD.108 There is some evidence that the combined application of perfusion MRI (adenosine stress and rest) and delayed enhancement MRI is associated with better accuracy than is perfusion MRI alone. A recent prospective study investigating 100 consecutive patients (51% women) with suspected CAD who were scheduled for coronary angiography, determined for the combined approach findings for sensitivity, specificity, and accuracy of 89, 87, and 88, respectively, compared with 84, 58, and 68%, for perfusion MRI alone.109

Coronary magnetic resonance angiography (MRA) is a method for non-invasive evaluation of epicardial coronary arteries. Since initial reports on this method in 1987, ongoing development of the technology involved has led to enhanced special and temporal resolution. Nevertheless, accurate determination of the presence as well as the severity of coronary stenosis, especially in tortuosity and small arteries, remains difficult due to breathing and heart motion. At present, direct visualization of the coronary arteries is not comparable with CT angiography. A meta-analysis of 25 studies with data on 993 subjects reported sensitivity and specificity for CAD detection of 73 and 86%, respectively.110 This meta-analysis disclosed no major differences between subgroups based on technical or population characteristics.

In conclusion, although limited data are available for women, cardiac perfusion and stress MRI appear to represent a promising approach for enriching the non-invasive diagnostic tool, for women as well—particularly those unsuited for stress echocardiography. A recent consensus panel accordingly assigned a Class-II recommendation here.111 A further advantage of MRI lies in the possibility of characterizing plaque composition. Research data thus suggest that high-resolution MR imaging may provide valuable information on plaque stability and may therefore potentially play a role in prediction of plaque-rupture risk.

Assessment of coronary microcirculation

A fundamental problem in understanding the terms ‘false positive’ and ‘false negative’ in women arises from the fact that most diagnostic tools refer to epicardial obstructive disease. However, there is growing evidence to support the concept that, in women, microvascular dysfunction rather than epicardial obstruction may be the source of ischaemia and symptoms. In women included in the WISE study, coronary microvascular dysfunction was identified in half the women with angina in the absence of obstructive epicardial disease.112 Hence, in many cases of angina and normal epicardial coronary arteries in women, additional methods to assess coronary endothelial dysfunction and/or coronary microcirculation may substantially improve the understanding of disease.

There are several approaches to assess coronary blood flow. One is the TIMI (thrombolysis in myocardial infarction) frame count, which—although useful for comparative purposes—appears insufficiently accurate for flow quantification.11,113,114 Two other methods—i.e. intracoronary flow wire measurements using the Doppler principle, as well as intracoronary thermodilution—allow a more precise coronary flow assessment.115118 Despite the fact that epicardial coronary blood flow alone provides only indirect information for coronary microcirculation, functional testing of coronary vasoreactivity yields further useful information. Intracoronary acetylcholine assesses endothelium-dependent vasodilation. Whereas a vessel with intact endothelium responds with epicardial dilation, coronary endothelial dysfunction is mirrored by impaired vasodilator response, or even by epicardial constriction. This phenomenon predicts acute coronary events as well as long-term disease progression.119121 In women with angina in the absence of epicardial obstruction, testing with intracoronary acetylcholine is recommended by recent ESC Guidelines to exclude endothelial dysfunction and vasospasm, as well as to determine endothelium-dependent coronary flow reserve (CFR).91

Nevertheless, there is no single method allowing direct visualization of coronary microcirculation in humans. CFR, the ratio between resting flow to maximum flow after dilation, appears to be a useful tool to abrogate the function of the microvasculature and its ability to respond to dilatory stimuli—based on the fact that flow resistance is primarily determined by the microvasculature. A CFR less than 2 after stimulation with dilators such as adenosine or dipyridamole is considered abnormal, and indicates microvascular dysfunction.11

With advances in cardiac imaging techniques, determination of MBF by positron emission tomography (PET) has become an attractive non-invasive diagnostic alternative. As shown in the WISE study, it also allows determination of CFR.122124

Under-use of diagnostic procedures

Female gender has been identified as a predictor of less-frequent use of exercise testing.125127 Moreover, when stress test results are positive, women are less likely to be referred to coronary angiography than are men. Recent analyses of stable angina conducted by the Euro Heart Survey with enrolment of 3779 patients (including 42% females), corroborate these data, which evidence that women are less likely to undergo an exercise ECG (−19%) and less likely to be referred for coronary angiography (−41%).58 The authors pointed out in their discussion that stress testing had not been as frequently used for women owing to the high number of false-positive results. At the same time, only 25% of the women who did not receive an exercise ECG were referred to an alternative stress testing technique. These sex-specific differences in stress test utilization may result in underdetection of CAD in women and treatment only at later stages of the disease. Nevertheless, it is presently unknown whether the outcome of women could be ameliorated with more extensive use of diagnostic procedures.

Once the diagnosis of CAD has in fact been established, reported sex-specific differences in the referral to diagnostic procedures are minor and are apparently more significantly related to the older age of women with manifested disease.128


Sex and gender differences in clinical presentation and results of diagnostic testing must be taken into account in clinical practice. Multiple risk-factor analysis, as well as determination of the pre-test probability of CAD in combination with more aggressive diagnostic testing, may increase CAD disease detection, among women as well. For asymptomatic and symptomatic women, classification into low, intermediate, or high risk categories, perhaps on the basis of risk scores, should represent the initial step (Figure 1).

Figure 1

Diagnostic algorithm for detecting coronary artery disease in women with chest pain

In asymptomatic women, imaging is not recommended in the low risk group. In those women with low pretest likelihood of the disease, stress testing is associated with a high rate of false positives. In asymptomatic women with intermediate risk, the value of routine diagnostic testing is controversially discussed. Further search for subclinical disease could support decision-making: e.g. in the form of IMT or peripheral endothelial function. Due to the high negative predictive value, CCSs or, even better, CT angiography, represent promising measures in patients with low-to-intermediate pre-test probability of CAD (Figure 1).

For initial assessment of symptomatic intermediate-risk women, an exercise treadmill test can provide useful diagnostic and prognostic information. The high negative predictive value of exercise testing—which additionally has advantages of wide availability and low initial costs—underscores recommendations to perform ECG stress test as first diagnostic procedure, also for women, along with normal resting ECG. The combination of several parameters such as exercise capacity and heart rate, in evaluation of positive stress results, can considerably improve diagnostic accuracy in these women. A test with imaging is recommended if further clarification is warranted, or if abnormal resting ECG results are obtained. Owing to its higher sensitivity and specificity (Table 2), stress echocardiography has been suggested to be the more reasonable approach for initial assessment of CAD in intermediate-risk women. In women not suitable for stress echocardiography, other stress-imaging methods involving radionuclide myocardial perfusion or MRI should be performed. In either case, initial diagnostic testing by an imaging approach—by application of stress echocardiography, SPECT perfusion imaging, or MRI techniques according to local expertise and/or availability—should of course be recommended in women who are not able to exercise. The benefits of non-invasive testing are greatest in patients with an intermediate-to-high likelihood of CAD. Women with intermediate risk who have negative test results are unlikely to have CAD and usually do not require further testing. Inconclusive results, however, call for further investigation. Cardiac CT angiography may be a valuable tool for women with non-conclusive stress test results and with intermediate likelihood of CAD. For symptomatically high-risk women, a more aggressive approach involving coronary angiography may be justified.

View this table:
Table 2

Meta-analyses on the accuracy of diagnostic stress tests in women

Author, yearExercise electrocardiographyStress echocardiographyStress SPECT
Sensitivity, %Specificity, %Sensitivity, %Specificity, %Sensitivity, %Specificity, %
Fleischmann et al., 199885778764
Kwok et al., 1999617086797864
Beattie et al., 200381737769
  • Reprinted with permission from Shaw et al.129

Microvascular dysfunction comprising the clinical constellation of typical angina and evidence of ischaemia in the absence of epicardial obstruction plays a greater role in women than in men. In these situations, invasive testing of coronary vasoreactivity allows exclusion or verification of endothelial dysfunction and coronary spasm. Moreover, invasive or—more elegantly—non-invasive determination by PET of the CFR enables assessment of the functional status of the microvasculature.

In conclusion, a combination of several parameters and diagnostic procedures can considerably enhance CAD diagnostic accuracy in women. In future, improved understanding of sex-specific differences in pathophysiology, clinical presentation, as well as diagnostic procedures—together with advancement of diagnostic technologies—will hopefully narrow the gender gap.

Conflict of interest: none declared.


View Abstract