Epidemiology of Cardiovascular Disease in Women

INTRODUCTION

The incidence and mortality rates of acute myocardial infarction (AMI) are greater in males than in females in all the population registries for this disease^1-4 (Figure 1). The AMI mortality ratio between males and females depends on age⁵ and varies between countries with a magnitude of 2 to almost 6 in the 35- to 64-year-old age group⁶ (Figure 2). On average, females who develop AMI do so 7 to 10 years later than males.^7,8 Despite these differences, it is believed that coronary heart disease will continue to be the leading individual cause of death in developed countries and, probably, in developing ones.⁹

Figure 1. Age-adjusted rates per 1 000 000 population in males and females at death due to ischemic heart disease in several developed countries in 2002.

Figure 2. Male/female ratio regarding death rate in 38 MONICA-WHO centers ordered by increasing population mortality rate in 35- to 64-year-old males. Adapted from Chambless et al.⁶

The advantage of females regarding incidence and mortality is lost when presenting AMI, since population mortality at 28 days is greater in females, especially in hospitalized patients.^10-36

Older age and the prevalence of comorbidity (in particular diabetes, hypertension and heart failure) in females explains some of these differences⁷ and have been cited as among the causes leading to this unfavorable situation. Nevertheless, the differences are maintained in many studies despite adjusting for these factores.^{10,11,14,15,19,30,32,33,36} Prognosis in the medium- and long-term is, however, similar in both sexes among survivors at 28 days from symptom onset when differences regarding the characteristics of both sexes are taken into account.^32-46

The problem of cerebrovascular disease (CVD) in Catalonia and Spain has continued to decrease in magnitude since the 1950s in terms of population mortality, and has done so faster than for ischemic heart disease. This fast reduction in CVD mortality means that Spain is among the countries with the lowest rates in the developed world (Figure 3). The standardized cumulative incidence rate in the 45- to 84-year-old age range in Catalonia only (268/100 000) (unpublished data) is slightly higher than that observed in the mid-1980s in France (238/100 000), and much lower than that in some developed countries.⁴⁷ The incidence and mortality (unpublished data) rates of CVD also are higher in males than in females⁴⁸ (Figure 4).

Figure 3. Rates per 100 000 inhabitants adjusted by age at mortality due to cerebrovascular disease in several industrialized countries in males and females, in 2002.

Figure 4. Specific (A) and standardized (B) death rates by age and sex per 100 000 inhabitants, due to cerebrovascular disease in those more than 24 years old by decade in Catalonia 2002.

Differences between sexes in the prognosis of AMI have been analyzed in observational studies and as a secondary endpoint in clinical trials as well as in other research that had not initially been designed to address this issue. All the approaches have advantages and drawbacks: whereas the population studies offer a broader vision that includes prehospital deaths due to these causes, hospital registries offer the opportunity to analyze in greater detail the clinical characteristics of the patients and comorbidity. A serious limitation that makes it difficult to compare the results of different studies lies in the fact of using different criteria to select patients.

MORTALITY FROM ACUTE MYOCARDIAL INFARCTION

Population Mortality

Population registries have the advantage of including patients who die from AMI before being admitted to hospital and, thus, they offer the opportunity to analyze mortality occurring before and after hospitalization in this population.

It has been found that fatal cases are distributed differently by sex: whereas sudden death more frequently occurs in males, females have a worse overall prognosis within 28 days after symptom onset with most fatal cases tending to occur in hospitalized patients and 24 h after admission.^6,22,24

This distribution indicates that females die more often from heart failure than from acute complications due to myocardial ischemia, such as ventricular arrhythmias.

Overall, age-adjusted mortality in females aged 35 to 64 years old is only slightly greater than in males (51.3 and 49.4%, respectively),⁶ although there is considerable geographical variability: in 13 of the 29 centers included in the MONICA study (Monitoring Trends and Determinants of Cardiovascular Diseases), the female/male ratio was significantly >1 but, surprisingly, in the remaining ones, no significant differences were found which were unfavorable to males (Table 1). In Spain, an interaction between sex and age regarding 28-day mortality has been described, such that females <64 years old do not have a worse prognosis than males of the same age, although those between 65 to 74 years old do.⁴⁹

Most fatal events (median, 70% in males and 64% in females) occur before the patients manage to get to hospital.^4,6,13,49-51 After admission, age-adjusted mortality is greater among females (26.9 and 21.8%, respectively; ratio, 1.24).^4,6

There is a strong inverse correlation between the population event rate and the mortality ratio between males and females (Figure 5). Southern European countries constitute an example of this phenomenon: there is a low incidence of AMI together with a high ratio of female/male mortality.^2,6 The reasons for this considerable geographical variability in population mortality can clearly be found in cultural differences, health systems and, no doubt, the actual differences in the incidence and severity of AMI. Despite all this, none of these hypotheses have been deeply explored up to the present, and should be an area of interest for future research.

Figure 5. Female/male mortality odds ratio in 38 MONICA/World Health Organization centers, ordered by population incidence of myocardial infarction in males from 35 to 64 years (A), and by the population incidence in females of the same age (B).

Observations in Hospitalized Patients

The high prehospital mortality in both sexes is a serious challenge to public health and if this can be reduced it would have a greater impact on total mortality at 28 days due to AMI than any other therapeutic advance to date. This stated, it has to be admitted that the health system also needs to direct its efforts toward the patients who manage to get to hospital and, thus, decrease hospital mortality due to AMI.⁵¹

There is a lack of detailed information on the clinical picture and medical history of a substantial number of patients who die before arriving at hospital, thus making it difficult to accurately determine the etiology of coronary death,^1,5 and who are later included in the population registries (1%-51%). However, this information is available regarding the patients included in hospital registries,^{6,12,21,49,52} which are, therefore, the only ones in which it is possible to determine whether the greater mortality in females after AMI is related to greater comorbidity or severity of the disease.²²

Table 1 presents the characteristics and basic results of the studies published to date where it was possible to estimate the relative risk of death after an AMI for females and to determine for which variables risk was adjusted. In 14 of the 19 registries which included patients consecutively, the relative risk (RR) for females was >1.20, and in 9 this was >1.39. In 10 of these studies, the RR was statistically significant. It needs to be emphasized that RR was 1.50 in all the studies on the Mediterranean area--mainly Spanish--and which included a broad age range. Only 3 of the studies reported RR less than 1; in 2 of them patients were >64 years old, but risk did not reach statistical significance.

In the MONICA-WHO study, a similar pattern was found in the relationship between the male/female 28-day mortality ratio and the incidence of AMI among hospitalized patients.⁵ The female/male mortality ratio was also higher in the areas with lower AMI incidence rates.^13-16

LONG-TERM MORTALITY

Few studies have compared male and female mortality beyond 6 months. Table 2 presents a selection of articles that described patient evolution between 6 months and 14 years.^{13,30,31,34,36-38,41,42}

If we consider the average age at the time of AMI symptom onset, a follow-up of more than 15 years would give similar results in both sexes.

A greater risk of mortality in females was found in follow-ups <1 year, even after adjusting for age and comorbidity.^10,14,39 In contrast, in studies which analyzed longer periods, no differences were observed between males and females,^{13,29,30,32-34,37,43,44} greater risk of death in females³⁵ or statistically significant lower mortality among females.^{31,36,38,41,42,44}

SOURCES OF VARIATION THAT AFFECT COMPARABILITY OF RESULTS

Selection criteria vary from one study to another. The population basis can change how the findings are viewed.⁵³ The upper age limit is one of the crucial factors for the assessment of population differences between sexes; however, the inclusion of incident cases or incident cases plus recurrent ones, non-Q wave AMI or patients with unstable angina in the hospital registries can also limit comparability between them.

Many studies that have addressed the role played by sex in AMI survival were not originally designed with this aim, such as the clinical trials or AMI registries which included non-consecutive patients.^{11,19,20,25,32}

There are a large number of follow-up times used to establish mortality: at 28 days at the population scale or in hospitalized patients, at 28 days among 24-h survivors, and prehospital mortality or mortality at 24 h, are all examples of the variability that can be found in the literature, together with other less precise ones, such as the period of hospitalization.

Another source of variation and uncertainty is related to some studies which excluded patients who died in the emergency ward. This bias is particularly important, since most deaths occur in the first 24 h.⁶

Tables 1 and 2 include the adjusted RR or odds ratio (OR). However, there is great heterogeneity in the number and type of variables included in this adjustment. Apart from age, which is clearly a confounding factor related both to mortality and sex, other variables related to previous risk in each patient also require adjustment in order to take into account their capacity to respond to the disease. Furthermore, the revascularization procedures used soon after symptom onset can radically change prognosis and should also be included in the models. Finally, to determine if the higher risk in females is attributable to greater severity, models can be included with variables such as cardiogenic shock, pulmonary edema or serious ventricular arrhythmias that could help to evaluate this situation of increased serious risk. Unfortunately, the multivariate analyses carried out in many of the studies listed in Tables 1 and 2 consist in step-by-step logistical regressions only. This fact hinders the comparability of results.

POSSIBLE EXPLANATIONS OF WORSE SHORT-TERM PROGNOSIS IN FEMALES AFTER A FIRST MYOCARDIAL INFARCTION

Killip class measures the presence and severity of left ventricular dysfunction and is one of the most powerful predictors of mortality after an AMI.⁵⁴ Females who present this have a more frequent background of heart failure than males and usually receive more diuretic and inotropic medication.²⁶ However, overall, females receive less treatment than males (see below).^55-59 Diastolic function during myocardial ischemia is probably related to the greater frequency of Killip class III-IV found in females in the acute phase of AMI. However, this is not necessarily accompanied by a worse ejection fraction (in fact, the opposite has been observed) or by more extensive necrotic lesions than in males.^28,29,35 As mentioned, females present worse Killip class than males during the acute phase of AMI.^13,28,59 Regardless of age at presentation, females develop more serious complications than males in terms of heart failure and reinfarction, even when ventricular function is similar at admission. This could indicate that there is a smaller cardiac reserve in females leading to worse diastolic function.²⁵ These possible differences between sexes regarding diastolic function probably require in-depth study.

Surprisingly, females also seem to develop mitral regurgitation more frequently, septal rupture, free-wall rupture, ventricular aneurysms, asystole and advanced atrioventricular block than males after AMI,^14,31,59 but less fibrillation or ventricular tachycardia.¹⁴

The possibility that females have smaller caliber coronary arteries, fewer collateral vessels or longer-lasting ischemia have also been suggested as explaining these differences.^21,25,28

Some theories are based on physiopathology. These include the existence of hypercoagulability states⁶⁰ and coronary arterial spasm,⁶¹ which are mechanisms described in young females that could explain greater mortality after AMI compared to males, both in the short- and long-term; such differences were not found when ages were >75.^62,63

A possible genetic mechanism has also been described, whereby females would be more susceptible to presenting ischemic events compared to males when there is a family history of ischemic heart disease.⁶⁴

PRESENTATION OF INFARCTION SYMPTOMS IN FEMALES

Some studies have shown that females present silent heart attacks more frequently than males after 55 years of age,²⁴ which could be easily explained by the greater prevalence of diabetes among AMI patients. This would also explain the fact that females present signs of serious heart failure as a first symptom of AMI more frequently than males.^13,16,21 It seems that females not only present more moderate symptoms of AMI, but more frequently develop atypical symptoms, such as abdominal discomfort and dyspnea.⁵⁴ Between 13 and 25% of myocardial ischemic episodes lack symptoms due to the presence of diabetes and older age.²⁸

DELAYED HOSPITALIZATION

On average, hospital admissions are delayed by 1 h compared to males, probably due to the atypical symptoms.^56,59 This factor, together with those described above and older age, would explain the lower use of thrombolysis, and, partly, the worse short-term prognosis.^{14,25,28,29,52}

USE OF DIAGNOSTIC AND THERAPEUTIC PROCEDURES

Females also receive less aggressive drug treatment, with less aspirin, beta-blockers (both in the acute phase and at discharge) and angiotensin-converting enzyme inhibitors.^56,58,59 Such differences are probably explained by older age, comorbidity and Killip class at admission. Furthermore, females undergo fewer diagnostic procedures (coronary angiography) and therapeutic ones (such as coronary artery bypass graft surgery and angioplasty) and these are done later than in males, even after adjusting for age and AMI location.^{11,12,25,54,55} In some studies it seems that the percentage of coronary angiographies and percutaneous interventions is lower among females, but this difference disappears after adjusting for comorbidity and age,⁵⁹ and only persists in cases where the indications are more uncertain.⁵⁷ A recent substudy in patients with non-ST elevation acute coronary syndrome highlights the underuse of angiography in females, especially in high-risk groups, as well as a short-term increase in refractory angina and readmissions due to angina.⁵⁸ If it is taken into account that females with AMI present more serious symptoms than males, it is reasonable to assume that even the absence of differences in the use of diagnostic procedures and invasive treatments can be interpreted as their underuse in patients who would benefit from a more aggressive approach.⁵⁸ In the countries with a low incidence there are no observed differences in the use of these procedures.^13,14

Given the difficulties encountered when comparing the results of published studies, it seems advisable to find a way to analyze and present standardized results, which could consist of including consecutive cases of Q-wave AMI admitted to hospital (not only those with a coronary care unit). Neither is it necessary to impose age limits, but it is advisable to carry out subanalyses in the 25- to 74-year-old subgroup. Standard follow-up at 28 to 30 days and adjusting for risk of death in females by age, diabetes, hypertension, and smoking are equally recommended to facilitate comparability between studies.

Overall, mortality due to CVD is lower in females up to 84 years old, and population mortality due to AMI is from two to seven times less than in 25- to 64-year-old males. This advantage is lost once a first AMI has occurred: mortality at 28 days in females tends to be around 20% greater when adjusted for age, especially among those from areas with a low incidence of this disease. In hospitalized patients, mortality is greater in females, but exclusively among patients with a first Q-wave AMI: this difference has not been found in Spain for the remaining acute coronary syndromes.⁵² There is a perceived difference in the distribution of deaths between males and females during the 28 days from symptom onset that indicates different death mechanisms: ventricular fibrillation in males and ventricular failure in females. The treatments used are proportionally less aggressive in females. All this indicates that a change of attitude is needed in all health contexts so that it is possible to more promptly identify females who have begun to show symptoms of an acute coronary syndrome in order to accelerate diagnosis and increase the use of diagnostic and therapeutic procedures, such that they are proportional to the severity of the picture presented.⁶⁵

---

Section Sponsored by the Dr Esteve Laboratory

Correspondence: Dr. J. Marrugat.

Unitat de Lípids i Epidemiologia Cardiovascular. Institut Municipal d'Investigació Mèdica.

Dr. Aiguader, 80. 08003 Barcelona. España.

E-mail: [e-mail]