Clinical Review:
Noninvasive Diagnosis of Coronary Artery Disease in Women

Carolyn M. Kosack,, M.D.,
The Hoffman Heart Institute of Connecticut
Assistant Clinical Professor of Medicine
University of Connecticut Health Center

In this issue of the Journal, we will discuss another non-invasive modality of testing for coronary artery disease and examine the gender bias that has been present in the evaluation of chest symptoms in women.

Stress Echocardiography

Few data are available on the accuracy of stress echo in women. Studies done in males showed the sensitivity of exercise echo to be 61�86% and the specificity to be 75 percent to 100 percent in a population where the prevalence of CAD was high. The value of exercise echo for detection of CAD in women was initially explored by Sawada et al. This study looked at 57 women with a 49% prevalence of CAD who underwent exercise echo for the evaluation of chest pain. They concluded that exercise testing with two dimensional echocardiography was useful in the diagnosis of CAD in women, with exercise echo having a sensitivity and specificity of 86 percent. They also found that this sensitivity was maintained in the presence of single vessel disease which is thought to occur more frequently in women.

Masini et al in Italy, looked at high dose dipyridamole echo compared to exercise echo in assessment of CAD and found no difference between the two in regards to sensitivity (79 percent vs. 72 percent) and the negative predictive value, (84 percent vs. 64 percent). The specificity and positive predictive value however were greater with dipyridamole echo. No studies utilizing dobutamine stress echo have been performed in females.

Other studies have explored the relative value of exercise echo and perfusion studies utilizing SPECT in identifying the presence and severity of CAD. Salustri et al compared exercise echo to perfusion SPECT (Tl-201 & Tc99m sestamibi) in 44 patients with stenosis of a single vessel. This paper concluded that in patients with single vessel CAD there was a high concordance 79 percent, between exercise echo and exercise SPECT imaging. The diagnostic accuracy for detecting CAD (more than 50 percent stenosis) was 72 percent to 86 percent for exercise echo and 75 percent to 77 percent for SPECT. Pozzoli et al examined the use of exercise echo and Tc99m sestamibi imaging in115 patients (68 men and 23 women) with suspected CAD and a normal resting ECG. Exercise echo identified 71 percent and SPECT 84 percent of patient without significant CAD with a concordance of 88 percent. This study also confirmed that the diagnostic accuracy of exercise echo and SPECT were significantly higher than that of exercise ECG alone, 81 percent vs. 64 percent and 88 percent vs. 64 percent, respectively. The sensitivity and specificity for detecting individual diseased coronary arteries were similar for both methods (60 percent and 95 percent for exercise echo and 67 percent and 94 percent for SPECT). However, in a subgroup of 33 patients with single vessel disease exercise echo had a lower sensitivity than SPECT, particularly for isolated stenosis of the left circumflex artery.

In summary, exercise and pharmocologic stress echocardiography may be very useful in female patients. However, at this time there is a paucity of published data on the accuracy of this modality and no information on the prognostic value of this method. Stress echo like MPI offers improved accuracy over the standard exercise ECG. Reports also suggest that the use of echo, added to exercise or pharmocologic stress testing, results in similar sensitivity and specificity as that seen with nuclear imaging. Initial referral center studies have included women with a higher prevalence of CAD and there is a need for more studies in a generalized population. Validation of dobutamine stress echo also needs to be studied further.

Gender Bias

Gender differences in the evaluation and management of patients with suspected CAD was first reported by Tobin et al in 1987. They noted that the referral rates for cardiac catheterization for men and women with abnormal nuclear stress tests were markedly different 40.2 percent vs. 4.2 percent. Men were more than 6.5 times likely to be referred even after controlling for age, prior MI, abnormal test results and degree of angina. A partial explanation for this was thought to be a result of clinical decision making by the primary care physician. The clinician, when faced with abnormal results from tests which historically were less accurate for women (who also had a lower prevalence of the disease), had a higher threshold for subjecting the patient to an invasive procedure with potential risks.

More recently in 1991, Ayanian reported on 80,000 male and female patients in Massachusetts and Maryland and found that men were twice as likely to be referred for angiography or revascularization procedures. Krumholz in a smaller study found referral for catheterization to be similar in men and women hospitalized for MI. This suggests that once the diagnostic dilemma is eliminated by clinical evidence of CAD (i.e. MI), the use of invasive tests to define extent of CAD equals that of men.

The Survival and Ventricular Enlargement Study (SAVE) was a study in which patients, 17% of whom were women, with MI and LV dysfunction were randomly treated with angiotensin converting enzyme inhibitor. This showed that despite greater functional disability and equal rates of angina prior to their MI, women were less likely to have been referred for catheterization than men (15.4 percent vs. 27.3 percent), strongly suggesting that clinicians may minimize anginal symptoms in women and are reluctant to proceed with more definitive diagnostic tests.

There appears to be significant evidence for gender bias in the diagnosis of CAD in women, further reinforcing the need for better noninvasive testing for this diagnosis in women. In 1994, Shaw et al evaluated a cohort of 3975 middle aged patients (840 women � 47 percent of the group), who were evaluated non-invasively for suspected CAD and they measured the rates of subsequent diagnostic procedures, coronary revascularization, MI or death. Hypertension, hypercholesterolemia and atypical angina were more common in women. Rates of initial positive tests (exercise induced horizontal or downsloping ST depression more than 1mm or more than 1 reversible thallium defect) were similar in women and men. Compared with men, most women with an initial positive test result had no additional evaluation, 62.3 percent vs. 38.0 percent. Coronary revascularization, not surprisingly, was done more frequently in men 4.9 percent vs. 2.0 percent for women. Cardiac death or MI occurred more often in women during 2 years of follow up, 6.9 percent vs. 2.4 percent for men. The decreased use of further diagnostic studies may have placed these women at greater risk for a subsequent adverse outcome, suggested by the higher cardiac event rates in these women. Finally, this study indicates that men and women share few common clinical predictors for increased rates of subsequent diagnostic testing and adverse cardiac events. However, for both women and men the single greatest outcome predictor was the presence of an abnormal non-invasive stress test or abnormal myocardial imaging study. In 1996, Hachamovitch et al analyzed rates of catheterization among patients who had undergone risk assessment by nuclear testing. They found that early catheterization referral rates (within 60 days), after nuclear testing in men and women was a function of the scintigraphic summed stress score. Fewer than 2 percent of patients with normal scans were referred for catheterization as compared to 16 percent of patients with mildly abnormal scans and 40% of patients with severely abnormal scans. The referral rate was not significantly different for men and women. This illustrates that nuclear testing may strongly influence subsequent patient management.

Recommendations

The best recommendations that can be made considering all the data we have to date on the diagnosis of CAD in women are as follows: