Gender and Coronary Disease

Nearly a decade has passed since the relation between gender and coronary disease became a hot topic. During this decade, articles have appeared describing the differences between women and men regarding the presentation of acute myocardial infarction (MI), the prognosis after an MI, and the use of interventions for diagnosing and treating coronary disease. These articles have contributed to the impression that women with coronary disease have worse outcomes than men.

The apparent difference between women and men may be explained by three complementary hypotheses:

• The pathophysiology of coronary disease is different in women. Coronary pathologyinteracts with the biologic sex characteristics of women to produce different symptoms and worse outcomes after acute events.

• Female sex, more than male sex, is associated with confounding factors that influence the diagnosis or prognosis of coronary disease. A woman and a man with othersie similar personal and clinical characteristics, however, have similar manifestations and outcomes of their coronary disease.

• Women and men with coronary disease are treated differently and this difference may be appropriate or due to bias.

There is a complex interplay among these potential explanations. The relative protection against coronary disease in women conferred by estrogen represents a fundamental difference that is related to a biologic sex characteristic. This protection is incomplete and wears off with increasing age, especially after menopause. The epidemiologic implications of this difference are important to understanding the observed gender differences in coronary disease. Although women develop coronary disease less often and at a later age, usually lagging behind men by about 10 years,¹ coronary disease is still the leading cause of death among women.² Also, older age is generally a poor prognostic factor and alters the expression of many diseases, including coronary disease. It is now well established that the older average age of women with acute MI (as well as the greater frequency of other, adverse, prognostic factors in women) accounts for the higher observed mortality in women after MI.³ Similarly, the increased mortality in women after coronary bypass surgery is associated with their older age and worse functional status.⁴

It is accurate to characterize women with coronary disease, collectively, as older and sicker than men with coronary disease. Therefore, it may not be inappropriate for there to be some differences in the average treatments they receive.⁵ Despite these differences, however, there is no justification for managing an individual woman with suspected or known coronary disease differently than if she were a man with an identical clinical picture. The tendency to under-diagnose and under-treat women with coronary disease, especially when symptoms are mild or the evidence of coronary disease is equivocal, was made famous as the “Yentl syndrome.”⁶ This tendency may be most insidious early in the course of coronary disease when patients first present with symptoms.⁷

In this issue, Zucker and colleagues report gender differences in the presentation of acute coronary disease.⁸ This issue is important because it is useful to identify patients with acute MI promptly while they are still eligible for thrombolytic therapy, and women receive thrombolytic therapy less often than men, even after adjusting for differences in age and eligibility criteria.^9,10 Zucker et al. present data about a large and diverse group of patients from multiple sites who are generally representative of women and men presenting to an emergency department with suspected MI. Since women are more likely than men to have acute MI without chest pain,¹¹ this study is noteworthy because it included such patients.

Among patients with suspected MI, women had significantly fewer MIs than men in all age groups, with rates approximately one-half those of men. These results are similar to those from a study of patients with chest pain.¹² Despite their overall differences, men and women with two findings during the initial evaluation had similar probabilities of MI. The presence of classic electrocardiographic changes, such as ST-segment elevation, was as specific for MI in women as it was in men, although this finding was seen in only a small proportion of patients. This observation confirmed the results of an earlier study.¹² Also, women with congestive heart failure (CHF) were as likely as men with CHF to have infarction, although only a minority of patients had CHF, and it is not clear whether CHF was new, worse, or old.

Why then, among the majority of women and men with suspected MI but without classic electrocardiogram changes or CHF, did fewer women “rule in” for an MI? The answer appears to be related to the lower prevalence of coronary disease in women. If symptoms are considered diagnostic tests and their sensitivity and specificity for MI are the same in women and men, the positive predictive value for MI still will be lower in women because women have a lower prior probability of coronary disease. This rationale also explains why women with typical anginal symptoms less often have confirmed coronary disease at angiography ¹³ and why the positive predictive value of an exercise test is lower in women.¹⁴ Although there may be some physiologic differences in exercise test performance between men and women,¹⁵ and the Zucker study suggests that CHF may have had a different predictive value for MI when women were compared with men, most of the differences in the predictive value of symptoms appear to be related to the lower prevalence of coronary disease in women.

This explanation notwithstanding, the Zucker study raises additional issues. Despite their lower prevalence of coronary disease, women in the Zucker study came to emergency departments complaining of symptoms suggestive of MI in numbers equal to those of men. This finding confirms earlier research ¹² and stands in contrast to the popular perception that women are neither sufficiently aware of MI symptoms nor motivated to have their symptoms evaluated.

Although this study provides us with useful information, its results have few clinical implications. Among patients who “ruled in” for an MI because of strong evidence of infarction on the electrocardiogram, the MI rate in women was not significantly different from that in men. For the majority of patients with a nondiagnostic electrocardiogram, the absolute risk for MI in this study was 10% for men and 6% for women, and the absolute risk difference was less than 8% among all age groups. These differences probably are not clinically meaningful for the individual patient, even though the difference in the relative risk for MI was approximately twofold. There also was no evidence from this study that gender could be used to weigh the importance of symptoms or signs in suspected MI, except possibly those of CHF. Gender, by itself, is a poor diagnostic test for MI and should not be used as the basis for triage or treatment decisions.

Although this study does not provide information about triage and treatment decisions, there is evidence from other studies that women with symptoms of cardiac ischemia are admitted to the hospital less frequently than men ^7,12 and receive follow-up diagnostic testing less often.⁷ Conversely, among patients who are admitted to the hospital because of chest pain, fewer men have a final diagnosis of coronary heart disease,¹² and women having an acute MI are no more likely to be sent home mistakenly from an emergency department than men.¹⁶ It thus appears that overuse of medical resources in men with suspected coronary disease may be as much a problem as underuse of these resources in women.¹⁷

The study by Zucker et al. also illustrates a more fundamental difficulty with studies of gender differences in coronary disease. Like almost all other published studies, this study relied on data that were originally collected for some other purpose. Gender is a complex concept, and there are physiologic, behavioral, socioeconomic, and clinical factors that may confound its relation to coronary disease. For example, information on the women's hormonal status was not reported. Also missing is information about any delay in receiving acute care for possible MI. Several studies have reported that women with symptoms of MI present later after the onset of discomfort than men.^11,12,18 While delay has obvious implications for the effectiveness of therapy, it also may affect how signs and symptoms are measured and interpreted.

The challenge to people conducting research on the relation between gender and coronary disease is to design and execute studies that take into account the complex factors associated with sex differences. In the meantime, readers and editors should be wary of post hoc analyses of gender differences in the management of coronary disease, and clinicians should pay careful attention to any evidence of coronary disease in women, just as they do in men.