Sex Differences in Myocardial Infarction and Coronary Deaths in the Scottish MONICA Population of Glasgow 1985 to 1991
Presentation, Diagnosis, Treatment, and 28-Day Case Fatality of 3991 Events in Men and 1551 Events in Women
Background Scottish MONICA used medical and medicolegal records and World Health Organization MONICA Project criteria to register coronary events in 25- to 64-year-old residents of the high-incidence area of north Glasgow from 1985 to 1991.
Methods and Results Age-standardized data from 3991 episodes of nonfatal definite myocardial infarction and coronary deaths in men (mean age, 55.5 years) were compared with 1551 in women (57.0 years). Many results, such as the overall 28-day fatality rates of 49.8% in men and 48.5% in women, showed insignificant differences. However, 74.3% of deaths in men occurred out of hospital versus 67.8% in women (P=.0004). After admission to hospital, fatality rates in women were 14% higher (P=.07) and after admission to coronary care, 22% higher (P=.04). Women were more often widowed. Fewer had a history of previous myocardial infarction, but the prevalence of angina pectoris, of smoking, and of chest pain in the attack was the same as in men; more had shock, syncope, and breathlessness. More consulted a doctor before admission to hospital, which delayed their coming under care. More men had ECG Q-wave progression, and more women had smaller ECG changes. This, and marginally reduced chances of direct admission to coronary care, of thrombolysis, of aspirin, and of β-blockers, did not explain women’s excess hospital fatality.
Conclusions Acute coronary events appear to be recognized and treated fairly equally in men and women 25 to 64 years old in Glasgow, so differences are small but subtle. More men die suddenly out of hospital; the reason why more women die after arrival may be because the equivalent number of men have already died outside.
Acute coronary events (myocardial infarction and coronary deaths) receive less attention as a cause of morbidity and mortality in women than in men. Men have more coronary events in middle age. Proportionately more middle-aged women die of cancer, involving in particular the female organs, so that premature cancer is popularly identified as the threat to women’s survival, whereas coronary disease is associated with men. Yet coronary heart disease is the leading cause of death in middle-aged women.1
Recent studies of angina pectoris suggest unequal investigation and intervention in women compared with men,2 3 4 5 6 despite angina being more evenly prevalent in the two sexes than is the incidence of acute events.7 8 These findings are complicated by the different clinical spectrum and diagnostic probabilities in women,1 4 9 but they raise questions about myocardial infarction and coronary deaths. Are acute events similar in men and women? Are emergency treatment and investigation equal or unequal?
Previous attempts at comparison have encountered problems. Smaller numbers of events in women lead to imprecision. Older age and more risk factors confound the analyses,10 11 12 13 14 15 16 as may failure to recognize mild attacks, so that cases in women appear more severe.17 18 19 20 Results of studies have been inconsistent.21 22 23 24 25 26 27
The World Health Organization (WHO) MONICA Project, and in particular the Scottish center based in the north of Glasgow city, is well placed to explore these questions. Thousands of coronary events are being investigated and coded over a period of 10 years to investigate disease trends in men and women <65 years old.19 The study covers all recognized fatal and nonfatal events in defined communities, both inside and outside the hospital. To ensure the stability of registration of events over time, strict diagnostic criteria and quality assurance procedures are laid down, monitored from the WHO Quality Control Centre for Event Registration in Dundee, Scotland, which also coordinates Scottish MONICA. The Scottish MONICA center in Glasgow has the highest female coronary event rates of all the centers; has scored well for the quality, completeness, and consistency of its data within the international study; and has recorded medical care data on all events continuously from 1985, rather than just periodically, as the original WHO MONICA protocol demanded.
What follows is a comparison of acute coronary events in men and women 25 to 64 years old from the first 7 years of the project, 1985 through 1991. The setting, chosen for its high event rates, is an inner-city area with a high degree of social and material deprivation, served by several large university and other hospitals providing free emergency care through the Scottish Health Service. Compared with North America, the level of invasive interventions in myocardial infarction is low, but during the years concerned, drug therapy was changing rapidly through involvement in and implementation of large, randomized, controlled trials.
The study population at risk is all 25- to 64-year-old residents of Glasgow city north of the river Clyde, whether their attacks occur in Glasgow or elsewhere. After pilot studies in 1983 and 1984, to determine the best ways of finding cases, definitive registration of coronary events began for the calendar year 1985 from an office in Glasgow Royal Infirmary. Discharge codes of hospital admissions for Glasgow are obtained from the Greater Glasgow Health Board and for elsewhere in Scotland from the Information and Statistics Division of the Scottish Common Services Agency. Hospital records are extracted by data clerks for all cases of myocardial infarction and for emergency admissions for coronary heart disease, and the ECGs are photocopied. The diagnostic data are coded by trained project nurses, all procedures being subjected to internal and external quality control, including a high rate of duplicate sampling and coding. ECGs are Minnesota coded28 by two coders independently of each other and of the case notes, with a third coder to adjudicate discrepancies. Domiciliary cases were pursued initially but proved to be too rare to warrant continuing. Death certificates for relevant causes are obtained as computer printouts from the office of the Registrar General for Scotland, as are annual population data. Reports of police inquiries into sudden deaths, which include details of lifestyle, medical history, and medication, are obtained from the Procurator Fiscal’s office, along with autopsy reports. General practitioners complete a questionnaire on fatal cases, and further data are obtained from hospital accident and emergency and pathology departments.
Patients and relatives are not interviewed directly. The data are all obtained after the event, in a standardized manner, from written records at second hand, a method we have labeled “cold pursuit,” in distinction to the obtaining of data from inpatients, or “hot pursuit,” practiced in earlier British studies and elsewhere.29 30 Register procedures and access to data were approved by the relevant committee. Personal identification needed for record linkage is stored on a computer separate from that used for medical information.
Diagnostic criteria and coding rules are based on the WHO MONICA Manual,31 extended by many additional items in the Glasgow MONICA Manual of Operations (unpublished). Data items are subjected to range and logic checks and a diagnostic algorithm and are entered on the Dundee University computer, from which core data are sent for the international study to the MONICA Data Centre in Helsinki, where many checks, as well as the diagnostic algorithms, are rerun, reinforcing the local quality assurance procedures. Rigorous efforts are made to prevent and detect changes in coding over time and thereby to confirm that any apparent trends detected are genuine.
Each event is defined as potentially lasting no longer than 28 days. After that time it can be classified as fatal or nonfatal, and results of investigations accruing in that time are used for diagnosis. A recorded history of myocardial infarction more than 28 days previously in the same individual determines whether the index event is first or recurrent.
The major nonfatal end point for the study is “definite” myocardial infarction. Cases failing to meet the criteria are classified as “possible” (with prolonged pain of myocardial ischemia but test results short of infarction) or “no” myocardial infarction. For a case to be coded as definite nonfatal myocardial infarction, there must be unequivocal serial ECG progression (defined by the Minnesota codes) during the attack or specific combinations of symptoms and ECG findings, with cardiac enzyme levels exceeding twice the upper limit of normal. (The latter varied by hospital and over time but could be any or all of creatine phosphokinase, hydroxybutyric dehydrogenase, aspartate transaminase, or their isoenzymes.)
Fatal events are categorized as “definite,” “possible,” “unclassifiable” (if no relevant diagnostic information is available for or against the diagnosis), and “no myocardial infarction or coronary death.” Definite fatal events have had diagnostic confirmation in life or autopsy evidence of myocardial infarction or coronary thrombosis. Possible fatal events have chronic occlusive coronary disease at autopsy, a past history of coronary heart disease, or suggestive symptoms before death in the absence of evidence for a competing cause of death.
For this analysis, cases have been included that satisfied criteria for the main WHO MONICA diagnostic grouping, definition 1.19 This includes definite, possible, and unclassifiable coronary deaths as well as nonfatal definite myocardial infarction. In particular, it should be noted that nonfatal episodes that failed the diagnostic criteria for definite infarction (the nonfatal possibles), which make up a large proportion of coronary admissions, are not included.
The tables compare percentage frequencies of findings for the events in men with those in women. Because the age distribution of events in the two sexes is slightly different (Table 1⇓), the uncorrected percentages are not directly comparable, although they are valid within each sex. To avoid the need to recalculate figures for both sexes to another standard, the percentage figure for women alone is reported once in its uncorrected (crude) form and then a second time, standardized directly to the male age distribution, and displayed alongside the male percentage for direct comparison.
For all tables, unless specified differently, the denominators are 3991 events in men and 1551 events in women; these include all nonfatal and fatal events, including out-of-hospital deaths, first events, and recurrences. To the extent that individuals remaining <65 years old and resident in the study population had more than one coronary event in the period 1985 through 1991, these totals and the following descriptions are of different events but not necessarily of different people.
Log-linear models were used to test differences adjusted for age. Probability (P) values are shown in the tables for each item, to the right for isolated positive responses, but underneath where there are multiple related alternatives whose distribution is analyzed together. In each case, tests were also made for age modification of the sex effect, and the significant (P<.05) results are reported.
The large numbers and standardized methods mean that results are of sufficient precision to indicate which findings show close congruence between the sexes and which show differences large enough to be of both statistical and real significance.
Numbers by Age, Sex, Social Background, and Place of Onset
Numbers of acute coronary events by age and sex are shown in Table 1⇑, together with the numbers that were fatal within 28 days. The breakdown by 5-year age groups of the events in men is that used for subsequent age standardization of percentages in women. There is a sevenfold ratio of male to female events at age 25 to 34 years, decreasing to a doubling at age 60 to 64 years. The nonfatal events are restricted to those that satisfy MONICA diagnostic criteria for definite myocardial infarction, whereas the fatalities include those inside and outside hospital (definite, possible, and unclassifiable), so the 28-day community case fatality is high. The all-ages (crude) case fatality rate is almost identical in the two sexes, at 49.8% in men and 49.6% in women. Although age standardization in women changes the figure to 48.5%, the difference from men remains insignificant. In both sexes, community case fatality rises with age from 40 to 44 up to 60 to 64, but with much smaller numbers there also appears to be a higher case fatality in the youngest age groups. Case fatalities, including those for hospital and coronary care unit admissions, are considered further in Table 10⇓.
Table 2⇓ shows the marital and social status of those affected. There are large and statistically significant differences. There are three times the percentage of events in widowed women as in widowed men after age correction. Only one third of the events in men and one fifth of those in women occurred in those in employment when afflicted, but for this item the differences by sex increased with age. In this inner-city, deprived area, few families owned their own housing, and about one in six victims lived in single-person households. In almost all of these characteristics, the female victims were more socially disadvantaged and isolated compared with the men. More were widowed, living alone and in publicly owned accommodations.
The place of onset of the acute coronary attack reflects the activities of those affected. In >40% of events beginning outside hospital, the exact place of onset was unclear in the clinical records or the person was found dead, and no presumption could be made about whether symptoms had begun elsewhere. When it was recorded, more men had the onset while at work or in a public place, whereas rather more women were affected at home. In the small percentage of cases with onset in hospital (4%), the coronary event could be complicating another reason for the admission.
Medical and Smoking History
Previous history of coronary heart disease or stroke and smoking habits at the time of the attack are shown in Table 3⇓. Almost one in two of the men and women registered with attacks had some previous clinical diagnosis of coronary heart disease, in which angina pectoris also featured with equal frequency. Rates of previous myocardial infarction, however, were significantly different (P<.0001), with about one quarter of women having such a history versus about one third of men. A history of stroke was equally common in the two sexes.
The Glasgow base population has high (but falling) smoking rates in both sexes, linked to urban deprivation and a Glasgow history of tobacco trading and cigarette manufacture.32 For nearly one fifth of coronary victims, usually the fatalities, smoking information was unavailable from the medical and medicolegal records, but the proportion of known current smokers among all coronary victims was nonetheless still remarkably high and similar at 60% in both sexes, although significant differences by sex occurred among the percentage of ex-smokers and other smoking categories. What was smoked also differed qualitatively and quantitatively, since more men than women smoked noncigarette tobacco, and more of them also smoked ≥30 cigarettes a day.
Symptoms and Signs
Table 4⇓ shows the symptoms and signs of the attack, sudden death outside hospital accounting for almost all the unknown, absent data. The presence andduration of chest pain showed no difference betweenthe sexes, but women had a significantly greater prevalence of what the WHO MONICA Manualclassified as atypical, additional features. Althoughpercentages for some items were small, there was more left ventricular failure (or acute breathlessness), more shock, and more syncope in women. Sixty-threepercent of both sexes were classified as having typicalchest pain without other features, but more men hadno recorded information on their symptoms and more women were atypical.
Coming Under Care
Medical management is shown in Table 5⇓. Women are slightly slower in coming under care than men, but the difference only just reaches statistical significance (P=.04). Significant differences are apparent in the routes by which men and women came under care and in their subsequent management. Relatively more men tended to go directly to hospital, whereas more women tended to consult their family (general) practitioner first. For this reason, time to come under care was reanalyzed by the route of coming under care and confining the analysis to those who were still alive at 28 days, since fatal and nonfatal cases are combined in the main tables. Of those going directly to hospital, 61.1% of men and 59.7% of women were under care at 4 hours from onset, whereas the figures for those consulting a family practitioner first were 36.1% and 35.1%, respectively (not significantly different). In the Glasgow population, therefore, the extra delay in women in coming under care is explained by their preference for consulting their family physician. This also explains why a few more were initially managed at home compared with men and why slightly more were taken to hospital by ambulance (not shown in table) rather than arriving by other means.
Although the differences in percentage terms are not great, those for management shown in this table reflect importantly on the 28-day fatality rates associated with hospital and coronary care unit admission, discussed subsequently in relation to Table 10⇓. If we compare the overall total of events in men and women, 4% more women than men were managed in hospital, whereas the equivalent 4% of events in men were medically unattended deaths outside, since the victims died before medical help could arrive. Rather more events in men were managed initially in a coronary care unit and more in women on a general medical ward, but the eventual total percentages of men and women admitted to coronary care and the length of stay there were indistinguishable.
Investigation and Diagnosis
The MONICA diagnostic categories and the noninvasive investigations on which they are based are shown in Table 6⇓. Medically unattended sudden deaths outside hospital account for both the missing investigations and the male excess in these, whereas investigation of remaining cases appears to be consistent across the two sexes.
The percentage breakdown of MONICA diagnostic categories resulting from these investigations is similar for some categories but subtly different for others. There is a male excess of definite myocardial infarction in which ECG changes alone were unequivocal and diagnostic (development of new Q waves, or Q-wave progression with ST and T progression). This is more than compensated for in women by an excess of cases, otherwise known as subendocardial myocardial infarction, that satisfied MONICA criteria for definite myocardial infarction through lesser ECG changes accompanied by cardiac enzyme levels exceeding twice the upper limit of normal.
As is implicit in the case mix of MONICA definition 1,19 the possible and unclassifiable cases in the diagnostic groups were all fatal, whereas definite myocardial infarction includes both nonfatal and fatal cases. Some fatal cases could be diagnosed on ECG and cardiac enzyme data obtained before death. The first three subgroups of definite myocardial infarction (1.1, 1.2, and 1.3) in Table 6⇑ include both fatal and nonfatal cases, whereas the fourth subgroup plus the possible and unclassifiable cases are all fatal. Most of the 28-day fatalities in the two sexes arise within these latter groups, but the first three accounted for 165 (8.3%) of the deaths in men and 80 (10.4%) of those in women. Within these categories, the average male case fatality at 28 days was 7.6%, whereas in women it was consistently higher, averaging 9.3%, so the excess fatality in hospitalized women was not explained by their different distribution from men across ECG groups.
Drugs and Interventions
Data on medication and procedures were extracted from hospital case records, and in fatal cases the Glasgow MONICA Project staff was able to obtain details of previous medication from police and general practitioner records. These are the 7-year averages for 1985 through 1991 during a period of rapid change. Results in Table 7⇓ are, first, for all coronary events, showing the recorded medication and investigation at or before the onset; then for medical treatment and invasive investigation of the attack, the denominator now restricted to medically attended cases (omitting unattended sudden death); and the third section showing medication and planned procedures at hospital discharge of survivors. Numbers and percentages of positives are given, data being recorded as unobtainable (not shown) in only 1% to 2% of cases for almost all items. Female percentages have been age-standardized to the equivalent male group.
Before the onset of the attack, drug usage was rather higher in women for several items, most notably diuretics. Cardiac procedure rates were low, but coronary angiography, bypass graft surgery, and coronary angioplasty rates had been marginally higher in the men. During the event, drug usage was remarkably similar in the two sexes, most items showing no significant difference, although the key drugs studied in randomized, controlled trials in the 1980s—aspirin, thrombolytic drugs, and β-blockers—were all used marginally more frequently in men. Again, procedure rates are low, with generally marginal excesses in men.
At discharge, most drug items show no significant difference between the sexes. Of 12 drug classes, 6 were used more often and 6 less often in women. Rates for procedures continuing or scheduled at hospital discharge were low, with no significant differences between the sexes.
Cardiac Arrest, Attempted Cardiopulmonary Resuscitation, and Death
Analysis of the available data on cardiac arrest and resuscitation attempts is shown in Table 8⇓. Consistent with what has been described already, more cardiac arrests occurred in men outside hospital and more in women after arrival in hospital. Adding the percentages for arrests in and out of hospital in men gives the same total as that for women (52.2% versus 51.8%). Recorded attempts at resuscitation out of hospital were made in 27.0% of arrests in men and (age-standardized) 26.9% for women. In hospital, the percentages were 77.6% for men and (age-standardized) 73.4% for women. Most out-of-hospital arrests occurred at or around the onset of the attack, but a tiny minority occurred between hospital discharge and 28 days.
Certain details of fatalities are compared in Table 9⇓. Denominators for this table are totals of deaths at 28 days, and the age standardization is to the male distribution of deaths. Details of the place of death contrast with the place of onset shown in Table 2⇑. There are far fewer unknowns, but again, more events in men than in women are related to public places and to place of work. Some 25.7% of male deaths were attributed to hospital locations, whereas the proportion in women after age standardization was 32.7%, despite their greater delay in reaching hospital. Coding of this item was to the place of initial fatal collapse, not of certification of death. According to MONICA coding rules, therefore, unsuccessful resuscitation attempts could not change the recorded time and place of death.31
The proportion of witnessed deaths was similar in the two sexes, even though more women lived alone. This is explained by survival times. Despite difficulties of estimation and a considerable amount of missing data, significantly larger proportions of women than men lived more than 1 hour and more than 24 hours from the onset, whereas many more men than women were recorded as surviving <1 hour. This is reflected in the subsequent investigation of the deaths. The more sudden nature of deaths in men and the greater number occurring outside hospital meant that more male deaths were investigated by the medicolegal authority, although the proportion of deaths in the two sexes subjected to medicolegal or hospital autopsies seems to have been the same.
Case Fatality in the Community and by Hospital and Coronary Unit Admission
Earlier tables have compared the community perspective of coronary events in men and women regardless of whether they reached hospital. Table 10⇓ summarizes the overall results but also shows the 28-day case fatality rates in relation to hospital admission and coronary care.
Community case fatality rates at 28 days are the same in men and women, but more women reached hospital alive in the attack, so that 74.3% of male deaths occurred outside hospital versus 67.8% (age-standardized) in women (P=.0004). This explains why 67.3% of women reached hospital and were admitted versus 63.3% of men (P=.005). The 28-day case fatality of those admitted to hospital was 13% higher in women than in men after age standardization (24.1% versus 21.4%, P=.07), whereas outcomes for coronary care unit admission showed a 22% disparity (19.3% versus 15.9%, P=.04).
The proportion of first attacks is different in the two sexes, but this does not seem to affect the community case fatality result. For first attacks alone, the figure is 48.5% for both sexes, indistinguishable from each other and from those found in all events (49.8% in men and 48.5% in women). The influence of whether attacks were first or recurrent on other potential findings is beyond the scope of this article.
The better hospital survival rates of men compared with women appears to be associated with equal and opposite differential survival outside hospital before admission. Male sex is associated with sudden death out of hospital and female sex with delayed death in hospital, but the overall fatality rates from onset to 28 days are indistinguishable in this large, carefully assembled case series.
The argument over whether men and women are fundamentally different or are made that way by society and whether women are treated equally provides the background to this study. We have exploited a uniquely suitable database. The long-term objective of MONICA is to monitor cardiovascular disease as objectively and precisely as possible over a period of 10 years.33 We have used it to provide, as far as we could, high-quality, standardized data on coronary events in men and women.
Validity and Generalizability of Glasgow MONICA Population Findings
The Glasgow MONICA population is not typical of the United Kingdom, or even of Scotland, because it was chosen in the early 1980s as a high-coronary-mortality population in the west of Scotland. Comparison with other populations, such as Edinburgh to the east, suggests that much of the geographical gradient in mortality correlates with socioeconomic disadvantage.34 35 The area of Glasgow, although not typical of the country, shares many features with a number of postindustrial conurbations in the northern hemisphere, which grew on heavy industry in the smokestack era, then lost traditional employment, and are now starting new industries.
Glasgow is populated by a mixture of descendants of lowland Scots and Celts from the Highlands and Islands and from Ireland, just across the Irish Sea. The ethnic diversity of what was a major trading port includes substantial additions in this century of Ashkenazi Jews, Italians, Poles, Chinese, and South Asians. Although Glasgow MONICA was the only MONICA population in the first risk-factor survey to achieve sexual equality in its (high) prevalence of cigarette smoking,36 attitudes toward women remain fairly traditional. However, society is changing, and Glasgow medical school recruited equal numbers of men and women students from the early 1960s on.
The Scottish Health Service follows a pattern similar to that of the rest of the United Kingdom in providing medical care free at the point of delivery, based on domiciliary care delivered through family practitioners by office and home visits, and through hospital consultants providing in-patient care and specialist hospital clinics. All acute-care hospitals in Glasgow have coronary care facilities, but these can be overstretched so that, through nonavailability of beds or atypical presentation, patients can be admitted on arrival to the general wards. From mid-1990 on, near the end of the period reported, all emergency ambulances in Glasgow were equipped with semiautomatic defibrillators.37 Before that, mobile coronary care was not generally available.38 Angiography, angioplasty, and coronary bypass surgery rates in acute myocardial infarction are low in Britain, where they have not previously been considered to be of proven value, and the system of funding discourages expensive interventions. However, in keeping with the pioneering British role in this area, Glasgow cardiologists are in the forefront of carrying out and acting on the results from randomized, controlled trials of drugs.
An important characteristic of the west of Scotland in the 1980s was its high mortality rates for a number of diseases, particularly in women. In the cross-sectional comparison of coronary event rates in WHO MONICA centers, out of 38 centers worldwide Glasgow MONICA had the third highest male event rates from 1985 through 1987 and the highest female rates by a large margin.19 Data (awaiting publication) from 1988 through 1990 give it both the highest female and the highest male rates, because male rates in Finnish centers had fallen further.
The Glasgow MONICA population is well placed to provide large numbers of events in women for analysis. Could results be distorted both by the factors generating the high event rates and by the high event rates themselves? We think not. Men and women are evenly distributed across populations, and consistent differences between them should be apparent in our population. Correction is needed only for age. Glasgow may avoid bias present elsewhere. When a disease is underrecognized, it tends to be the more severe cases that are diagnosed. The WHO MONICA Project has shown that case fatality is higher for women in populations in which coronary disease rates are low.19 The event rate for women in Glasgow exceeds that of the men in southern European populations. The disease in women should be better recognized than elsewhere.
Missed and “silent” cases are beyond the reach of disease registers and can be identified only from longitudinal cohort studies with regular ECG follow-up. They make up a substantial percentage of events in men,39 and the Framingham study suggested an even larger percentage in women.7 In Glasgow, the community case fatality and other indicators suggest that the case mix in men and women is similar, although the proportion of missed cases in the two sexes can only be guessed at.
Once a case is picked up from the hospital discharge or death certificate, the Glasgow MONICA registration staff follows a highly standardized procedure for extracting and coding the information and for deriving the diagnosis. In particular, the Minnesota coding of the ECG is carried out by observers independently of each other and of the case notes. The MONICA ECG code and the MONICA diagnosis are provided by a computer algorithm.
Contribution to the Evidence on How Men and Women Differ
Our findings build on earlier studies but, by incorporating out-of-hospital sudden deaths, place them in a new perspective. For many characteristics in the two sexes, we have shown that there are no significant differences. Where we have found differences, we have been able to quantify these with precision. The differences are subtle but of considerable importance. The community perspective shows that the worse in-hospital mortality encountered by women compared with men is compensated for by advantages over men before arrival. Women do appear to have a rather worse prognosis in hospital, but this may be because the equivalent number of men have already died suddenly before reaching hospital.
Evidence that these findings on sex differences are not unique to Glasgow MONICA comes from data on the 38 WHO MONICA populations.19 In 29 of the 36 populations able to provide comparable data for men and women, the percentage of deaths in men that apparently occurred within 1 hour of onset was greater than the percentage in women; in 2 it was the same, and in only 5 was it smaller. The pooled data also show a higher proportion of fatal cases in women than in men with ECG and enzyme tests before death who survived long enough to come under care.
Although some studies have not reported higher fatality in women from coronary events,21 24 most have done so, particularly those that reported on hospital admissions. They have shown great variation in whether they attributed the worse outcome to age alone11 15 or to combinations of risk factors,10 12 13 16 26 or concluded that sex was a risk factor in itself14 17 18 20 22 25 27 either from an inherent difference between the sexes or from underrecognition of milder attacks in women. Differences in reported risk between men and women have been very considerable in different studies. By excluding patients >65 years old and by age standardization using 5-year age groups, we have shown that the excess mortality in women in hospital still remains significant, although it is not very great. In our data, we have found a significantly greater prevalence of shock, syncope, and left ventricular failure in women under care (Table 4⇑) but not enough to explain the excess of hospital deaths. We have also found a significantly greater prevalence of non–Q-wave infarction in women (Table 6⇑), but the excess female mortality extends across all of our MONICA ECG categories.
Are Women Different, Are They Treated Differently, or Do They Behave Differently?
In a country with many traditional, although changing, attitudes toward women, we have found possible evidence of differences from men in investigation and treatment but large numbers of categories in which there were none. Numbers of ECGs and enzyme tests recorded after admission were similar in men and women, as was most medication before, during, and after the attack; the proportion eventually admitted to coronary care; and the length of stay there. Relatively more men were admitted directly to coronary care, possibly as a result of different ECG findings on admission. Also, during 7 years of rapid change, marginally lower percentages of women were recorded as having received aspirin, β-blockers, and thrombolytics during the attack, although women had more of some other drugs. The more invasive investigations and interventions were more common in men, and in some cases this reached statistical significance. However, in most cases these procedures were so uncommon, even in men, that the differences were not significant or able to influence survival rates. Resuscitation was attempted equally commonly in arrests outside hospital but was recorded marginally less commonly with arrests in women in hospital. We do not know the clinical circumstances. Our findings on drug usage are consistent with those from Worcester, Mass,23 in showing small, nonsignificant differences in the same directions and show less potential bias than other studies from Britain.14 40 Revascularization procedures in myocardial infarction were too infrequent to test whether there are differences between the sexes, as has been investigated in Worcester.41
Women took longer to reach hospital during an attack, but the difference was small and was explained in Glasgow by their preference for consulting the family physician, whom they knew, in preference to going directly to hospital. This preference has long been recognized by medical sociologists.42
The picture presented in the Glasgow population is one in which acute coronary events are so common in both men and women that they are recognized and treated very similarly, and the case mix and severity mix appear to be similar. This case mix is determined partly by what cases come forward for medical care and are recognized and partly by the WHO MONICA criteria for myocardial infarction. Relatively more cases of nonfatal possible myocardial infarction (not included in WHO MONICA definition 1) occur in women compared with men. Large numbers of these events were registered in the Glasgow MONICA population but are not included in this article. A more inclusive set of diagnostic criteria would have lowered the case fatality in both sexes but more so in women than in men, giving women a comparative advantage.19 43 The more specific WHO MONICA diagnostic criteria for nonfatal infarction account for the community case fatality of 50% currently observed, compared with that recorded in the earlier WHO Myocardial Infarction registers of 35% to 40% when nonfatal possibles were included.43 44 They partly explain the high 28-day fatality rates in both sexes in hospitalized cases in Glasgow, but these still remain inexplicably higher than those from many reported hospital series and from other MONICA centers despite equivalent care.
The less typical symptoms and less classic ECGs in women may have delayed admission to coronary care, but some bias in the decision to admit from the emergency room to coronary care cannot be totally excluded. Although for some items women fared less well than men, these small differences in the management of women should not have made a significant contribution to hospital survival, since in the bulk of cases and in most respects their management was similar to that of men.
One significant item in the comparison of male and female victims is the greater social and material deprivation of women. The large number of widows in Glasgow, even in women below age 65 years, is related to the high premature death rates in both sexes, but the results in Table 2⇑ are supported by other unpublished data from Scottish MONICA and the Scottish Heart Health Study in showing a greater social gradient of coronary risk for women than for men.
Fate of Women Compared With Men
The conclusion of this study is that women below age 65 years hospitalized with myocardial infarction have a higher 28-day fatality rate than do men, even after correction for differences in age. They have a greater number of atypical features, take slightly longer to reach hospital, and have more subendocardial infarction, but none of these factors appear to account entirely for their worse mortality experience after admission.
The most plausible explanation that we have found is that women have a differential survival pattern. They appear to be relatively better protected than men from sudden death outside hospital, so that the excess of deaths in men occurs before admission and the equivalent excess in women after. Although it is impossible to prove that the female deaths in hospital are those deaths that would have occurred in men had they reached hospital alive, the agreement in overall 28-day fatality rates is quite remarkable.
The implications of these observations may be that we should be looking for out-of-hospital factors to explain what happens in hospital to men and women. The factors that protect against sudden death in a coronary attack need to be better understood to protect the men. Conversely, we also need to know whether properly applied coronary care is widening or narrowing the survival difference between hospitalized myocardial infarction in men and women. Here we have presented 7 years of pooled data during an era of rapid therapeutic change. The 10-year data collection from the Glasgow MONICA population, now coming to completion, will provide an answer on differential trends.
Scottish MONICA is funded directly by the Chief Scientist Office of the Scottish Office Department of Health. It has also been helped by the international coordination and contacts provided through the World Health Organization, by support for the quality control work in Dundee from the National Institutes of Health of the United States, and by the interaction with Finnish colleagues through the MONICA Data Centre in Helsinki and the Scottish-Finnish Health Agreement. The views expressed in this paper are those of the named authors and not necessarily of the funding bodies nor of the following, who have made valuable contributions to data collection and analysis in Glasgow and Dundee since 1983: K. Barrett, C. Bauwens, H. Bilkhu, C. Bowman, C. Brown, I. Crombie, J. Graham, M. Hastings, M. Irving, M. Kenicer, E. Kesson, W. Leslie, W. Millar, M. Mitchell, J. Palmer, M. Robb, M. Sharkey, M. Shewry, M. Thornton, W. Tunstall-Pedoe, and A. Urie.
- Received September 18, 1995.
- Revision received November 15, 1995.
- Accepted November 19, 1995.
- Copyright © 1996 by American Heart Association
Kee F, Gaffney B, Currie S, O’Reilly D. Access to coronary catheterisation: fair shares for all? BMJ. 1993;307:1305-1307.
Wenger NK. Gender, coronary artery disease, and coronary bypass surgery. Ann Intern Med. 1990;112:557-558.
Tobin JN, Wassertheil-Smoller S, Wexler JP, Steingart RM, Budner N, Lense L, Wachspress J. Sex bias in considering coronary bypass surgery. Ann Intern Med. 1987;107:19-25.
Khan SS, Nessim S, Gray R, Czer LS, Chaux A, Matloff J. Increased mortality of women in coronary bypass surgery: evidence for referral bias. Ann Intern Med. 1990;112:561-567.
Smith WCS, Kenicer MB, Tunstall-Pedoe H, Clark EC, Crombie IK. Prevalence of coronary heart disease in Scotland: Scottish Heart Health Study. Br Heart J. 1990;64:295-298.
Kennedy JW, Killip T, Fisher LD, Alderman EL, Gillespie MJ, Mock MB. The clinical spectrum of coronary artery disease and its surgical and medical management, 1974-1979: the Coronary Artery Surgery Study (CASS). Circulation. 1982;66(suppl III):III-16-III-23.
White HD, Barbash GI, Modan M, Simes J, Diaz R, Hampton JR, Heikkila J, Kristinsson A, Moulopoulos S, Paolasso EAC, Van der Verf T, Pehrsson K, Sandoe E, Wilcox RG, Verstraete M, von der Lippe G, Van de Werf F, and the Investigators of the International Tissue Plasminogen Activator/Streptokinase Mortality Study. After correcting for worse baseline characteristics, women treated with thrombolytic therapy for acute myocardial infarction have the same mortality and morbidity as men except for a higher incidence of hemorrhagic stroke. Circulation. 1993;88:2097-2103.
Wilkinson P, Laji K, Ranjadayalan K, Parsons L, Timmis AD. Acute myocardial infarction in women: survival analysis in the first six months. BMJ. 1994;309:566-569.
He J, Klag MJ, Whelton PK, Yuchang Z, Xinzhi W. Short and long-term prognosis after acute myocardial infarction in Chinese men and women. Am J Epidemiol. 1994;139:693-703.
WHO MONICA Project, prepared by Tunstall-Pedoe H, Kuulasmaa K, Amouyel P, Arveiler D, Rajakangas A-M, Pajak A. Myocardial infarction and coronary deaths in the World Health Organization MONICA Project: registration procedures, event rates and case-fatality rates in 38 populations from 21 countries in four continents. Circulation. 1994;90:583-612.
Greenland P, Reicher-Reiss H, Goldbourt U, Behar S, and the Israeli SPRINT Investigators. In-hospital and 1-year mortality in 1524 women after myocardial infarction: comparison with 4315 men. Circulation. 1991;83:484-491.
Goff DC, Ramsey DJ, Labarthe DR, Nichaman MZ. Greater case-fatality after myocardial infarction among Mexican Americans and women than among non-Hispanic whites and men: the Corpus Christi Heart Project. Am J Epidemiol. 1994;139:474-483.
Prineas RJ, Crow RS, Blackburn H. The Minnesota Code Manual of Electrocardiographic Findings: Standards and Procedures for Measurement and Classification. Boston, Mass: John Wright; 1982.
Tunstall-Pedoe H. Problems with criteria and quality control in the registration of coronary events in the MONICA Study. Acta Med Scand. 1988;suppl 728:17-25.
WHO MONICA Project. MONICA Manual (Revised Edition). Geneva, Switzerland: Cardiovascular Diseases Unit, World Health Organization; 1990.
Crombie IK, Smith WCS, Tavendale R, Tunstall-Pedoe H. Geographical clustering of risk factors and lifestyle for coronary heart disease in the Scottish Heart Health Study. Br Heart J. 1990;64:199-203.
Watt GCM, Ecob R. Mortality in Glasgow and Edinburgh: a paradigm of inequality in health. J Epidemiol Community Health. 1992;46:498-505.
The WHO MONICA Project. A worldwide monitoring system for cardiovascular disease. In: World Health Statistics Annual 1989. Geneva, Switzerland: World Health Organization; 1989:27-149.
Leslie WS, Fitzpatrick B, Morrison CE, Watt GCM, Tunstall-Pedoe H. Out of hospital cardiac arrest due to coronary heart disease: a comparison of survival before and after the introduction of defibrillators in ambulances. Heart. 1996;75:195-199.
Fitzpatrick B, Watt CGM, Tunstall-Pedoe H. Potential impact of emergency intervention on sudden death for coronary heart disease in Glasgow. Br Heart J. 1992;67:250-254.
Grimm RH, Tillinghast S, Daniels K, Neaton JD, Mascioli S, Crow R, Pritzker M, Prineas RJ. Unrecognised myocardial infarction: experience in the Multiple Risk Factor Intervention Trial (MRFIT). Circulation. 1987;75(suppl II):II-6-II-8.
Clarke KW, Gray D, Keating NA, Hampton JR. Do women with acute myocardial infarction receive the same treatment as men? BMJ. 1994;309:563-566.
Chiriboga DE, Yarzebski J, Goldberg RJ, Chen Z, Gurwitz J, Gore JM, Alpert JS, Dalen JE. A community-wide perspective of gender differences and temporal trends in the use of diagnostic and revascularization procedures for acute myocardial infarction. Am J Cardiol. 1993;71:268-273.
Cartwright A. Patients and Their Doctors: A Study of General Practice. London, UK: Routledge & Kegan Paul; 1967.
Tunstall-Pedoe H. Uses of coronary heart attack registers. Br Heart J. 1978;40:510-515.
Myocardial Infarction Community Registers. Public Health in Europe No. 5. Copenhagen, Denmark: World Health Organization Regional Office for Europe; 1976.