Mehilli J, Kastrati A, Dirschinger J, Pache J, Seyfarth M, Blasini R, Hall D, Neumann F, Schömig A. Sex-Based Analysis of Outcome in Patients With Acute Myocardial Infarction
Treated Predominantly With Percutaneous Coronary Intervention. JAMA. 2002;287(2):210-215. doi:10.1001/jama.287.2.210
Author Affiliations: Deutsches Herzzentrum (Drs Mehilli, Kastrati, Dirschinger, Pache, Blasini, Hall, and Schömig) and 1. Medizinische Klinik rechts der Isar (Drs Seyfarth, Neumann, and Schömig), Technische Universität, Munich, Germany.
Context A higher mortality risk for women with acute myocardial infarction (AMI)
is a common finding in studies that compare the postinfarction outcome of
women vs men. It is not clear, however, whether sex is an independent predictor
of death among patients systematically treated with aggressive reperfusion
and medical strategies.
Objective To assess the impact of patient's sex on outcome in a consecutive series
of patients with AMI treated with a reperfusion strategy largely based on
percutaneous coronary interventions.
Design, Setting, and Patients Inception cohort of 1937 patients (502 women and 1435 men) who were
admitted with a diagnosis of AMI to a tertiary referral institution between
January 1995 and December 2000.
Main Outcome Measures Mortality at 1 year after AMI.
Results Compared with men, women were older (70 vs 61 years; P<.001) and had known diabetes or hypertension more often. Both
men and women received essentially identical therapy with the majority of
patients (86%) receiving reperfusion therapy via percutaneous coronary interventions.
There were no significant differences in 1-year Kaplan-Meier death rates with
13.8% (68 cases) among women and 12.9% (184 cases) among men (unadjusted hazard
ratio, 1.06; 95% confidence interval, 0.80-1.39; P
= .70). After age adjustment, women had a lower risk of death (hazard ratio,
0.65; 95% confidence interval, 0.49-0.87; P = .004).
Conclusion Despite their more advanced age and greater prevalence of diabetes or
hypertension, women with AMI who were treated with a reperfusion strategy
largely based on percutaneous coronary interventions show a similar outcome
In the last 20 years, cardiovascular disease has caused more deaths
among women than among men.1 Furthermore, from
1979 through 1998, mortality from cardiovascular diseases in the United States
has steadily decreased in men but has remained relatively constant and even
increased in women.1
Acute myocardial infarction (AMI) is the major cardiovascular cause
of death in men and women. A considerable number of studies have focused on
sex-related differences in outcome in patients with this disease. Prevailing
evidence supports that short- and long-term mortality is higher in women than
in men.2 Less intense use of both pharmacological
and device-based therapies, biological factors, older age, and more severe
cardiovascular risk profiles have been used to explain higher mortality rates
among women.3 Important limitations of previous
studies have included restriction to selected subsets suitable for thrombolysis4- 10
and the inclusion of cohorts of patients in whom reperfusion strategies were
used relatively infrequently.11- 20
Randomized studies performed in the past few years have demonstrated
that percutaneous coronary interventions (PCIs) are more effective reperfusion
strategies than intravenous thrombolysis.21,22
Outcomes of primary PCI in patients with AMI have improved further with the
use of the glycoprotein IIb/IIIa inhibitors.23- 25
However, PCI is still used less frequently than thrombolysis in patients with
AMI.26 Data on differences in outcome between
women and men with AMI treated with PCI are limited and have been restricted
to selected subsets of patients.27,28
However, an important secondary analysis of a randomized study suggested that
a difference in outcome in favor of men was found only after thrombolysis
and not after PCI.29
The objective of the present study was to assess the impact of female
sex on the outcome of a consecutive series of patients with MI treated with
an aggressive reperfusion strategy largely based on PCI.
This study included all 1937 patients who were admitted to the Deutsches
Herzzentrum and 1. Medizinische Klinik rechts der Isar in the period between
January 1995 and December 2000 with a diagnosis of AMI. The diagnosis of AMI
was based on the presence of at least 1 episode of chest pain within the last
48 hours lasting 20 minutes or longer combined with either typical electrocardiographic
changes (≥0.1 mV of ST-segment elevation in ≥2 limb leads or ≥2 mV
in ≥2 contiguous precordial leads, presumed new complete left bundle-branch
block) or increase of creatine kinase and its mb isoenzyme twice above the
upper limit of normal. All of these patients were followed up according to
a prospectively defined protocol to which they had given their informed consent.
The decision about which treatment approach to use was made by the responsible
physician. Primary PCI was the preferred approach and the use of thrombolysis
was mostly dictated by enrollment in randomized clinical trials (78% of the
cases treated with thrombolysis). During the entire study period, a continuous
on-call service was operative around the clock to provide the patients with
AMI the option of a prompt PCI.
The collection of hospital data and performance of 30-day, 6-month,
and 1-year clinical follow-up in patients with AMI for research purposes was
approved by the institutional ethics committee.
The main outcome measure was all-cause mortality. Other outcomes of
interest were recurrent nonfatal MI and stroke. Diagnosis of recurrent infarction
was based on typical chest pain, new ST-segment changes, and an increase in
creatine kinase of at least 50% over the previous trough level in at least
2 samples reaching 240 U/L or higher. The diagnosis of stroke required confirmation
by computed tomography or magnetic resonance imaging of the head. During the
hospital stay, creatine kinase and its isoenzyme were determined immediately
after admission and at least daily thereafter. After discharge, the assessment
of clinical status was made by means of a telephone interview at 30 days,
a follow-up visit at 6 months or whenever dictated by patient complaints,
and a telephone interview at 1 year. Only 3.0% of men and 2.9% of women were
lost to follow-up.
Differences between men and women were assessed using a 2-sided χ2 or Fisher exact test for categorical data as appropriate, and the
Wilcoxon rank sum test for continuous data. Survival curves were constructed
by the Kaplan-Meier method with differences in survival assessed with the
log-rank test. Multivariable analyses were performed by using Cox proportional
models after confirming the validity of the proportional hazards assumption
by the Therneau function.30 We formerly assessed
the interaction between sex and age with respect to mortality. Potential confounders
were entered into models if they showed univariable differences between women
and men with a P<.10. All statistical analyses
were performed using S-Plus Version 4.5 software (Mathsoft Inc, Seattle, Wash).
All P values are 2-tailed. A P value of less than .05 was considered significant.
Baseline characteristics according to sex are listed in Table 1. Women were older and were also more likely to have systemic
arterial hypertension and diabetes. Smoking was encountered less frequently
among women. They were also less likely to have a history of previous MI,
bypass surgery, or PCI.
Infarct characteristics according to sex are also listed in Table 1. Compared with men, women presented
at the hospital with a longer delay from onset of symptoms and tended to have
a higher frequency of anterior infarction. However, among patients who received
thrombolysis or PCI, there was no difference in the time interval from admission
to treatment between women (median, 95 minutes; interquartile range, 70 and
150 minutes) and men (median, 95 minutes; interquartile range, 61 and 150
minutes) (P = .33). Killip class distributions were
Treatment characteristics according to sex are shown in Table 2. Coronary angiography was performed in an extensive and
comparable proportion of men and women. Reperfusion therapy was instituted
in an essentially identical proportion of women and men and PCI was the predominant
strategy (applied in 86% of patients). Two thirds of the patients also received
the glycoprotein IIb/IIIa inhibitor abciximab as part of reperfusion treatment.
After primary treatment, men and women were treated frequently with β-blockers,
angiotensin-converting enzyme inhibitors, and statins in similar proportions.
The 30-day event rates are shown in Table 3. No significant differences between women and men were evident.
In addition, revascularization of the infarct-related artery was required
in 6.2% of the women and 7.2% of the men (P = .48)
within the first 30 days after the primary treatment. In the small group of
patients who did not receive reperfusion therapy, 32% of whom were in Killip
class III or IV, mortality rates during the first 30 days were 14.3% among
women and 20.7% among men (P = .56).
Target vessel revascularization rates 1 year after initial treatment
were 19.3% for women and 20.9% for men (P = .45).
Cumulative mortality curves are shown in Figure 1. There was no significant difference in 1-year Kaplan-Meier
death rate (13.8% or 68 cases among women and 12.9% or 184 cases among men;
unadjusted hazard ratio [HR] for 1-year mortality, 1.06; 95% confidence interval
[CI], 0.80-1.39; P = .70).
We next focused on age-adjusted mortality according to sex because of
the substantial age difference between women and men. Age was entered into
the model as a nonlinear variable. After adjustment for age, female sex was
associated with reduced mortality (HR, 0.65; 95% CI, 0.49-0.87; P = .004). Figure 2 shows
the mortality risk of women and men as a function of age. No significant interaction
was found between sex and age with respect to mortality. Age-adjusted analyses
were performed for several prespecified subgroups (Figure 3). Women in these subgroups had a consistent reduction in
age-adjusted risk (Figure 3).
We constructed a multivariable Cox model for 1-year mortality including
all characteristics from Table 1
that differed with P<.10 between women and men
(age, hypertension, diabetes, smoking, previous MI, prior coronary artery
bypass graft, prior PCI, anterior MI, heart rate, and time to admission),
as well as sex itself. Female sex was independently predictive of a lower
mortality (HR, 0.67; 95% CI, 0.50-0.91; P = .01).
The most powerful risk factor was older age, with an HR of 1.69 (95% CI, 1.50-1.89)
for a 10-year increment. Other strong risk factors were a history of coronary
artery bypass graft (HR, 2.41; 95% CI, 1.62-3.59) and diabetes (HR, 1.80;
95% CI, 1.38-2.36).
In this study, we included all 1937 patients who presented at our center
for primary treatment of AMI during a 6-year period. To avoid bias, no selection
criteria were applied regarding demographic characteristics, clinical status
at presentation, time interval from symptom onset, and treatment approach
used. Twenty-six percent of the patients were women, about 90% of the patients
had emergency coronary angiography, and 86% underwent primary PCI as a reperfusion
strategy. Importantly, both women and men received essentially identical therapy
with no sex-based difference in the use of PCI or evidence-based concomitant
medical treatment (use of aspirin, β-blockers, angiotensin-converting
enzyme inhibitors, etc). Considering the high proportion of patients treated
with primary PCI, the results of this study should be regarded as representing
what might be expected from a systematic interventional approach in the treatment
of patients with AMI.
The main finding of this study indicates that women with AMI treated
with this strategy had a prognosis similar to that of men, despite their significantly
older age at presentation. This finding complements that of a previous study
of ours including noninfarct patients with coronary stenoses treated with
stenting31; together, these findings suggest
that, in general, women with coronary artery disease benefit from emergency
or elective PCI to at least the same degree as men.
A major strength of this study is the inclusion of an unselected and
consecutive population of patients with AMI. This was achieved by including
all the patients who sought primary treatment in our institution. In this
regard, our study is similar to previous studies that included patients irrespective
of the therapy received.11,12,14,18- 20
Other reports have been confined to selected populations with a specific reperfusion
In line with previous reports, women in this study differed from men in that
they were older, more likely to have diabetes, arterial hypertension, anterior
infarct localization, and they presented at the hospital with a greater delay
from symptom onset. On the other hand, more male patients had a history of
prior MI and coronary intervention. However, after adjusting for all these
differences, women were found to have a significantly lower adjusted risk
of death than men.
We included all patients admitted during the period between 1995 and
2000. Accounting for the period in which the studies are performed is important
for reducing the bias associated with recent treatment advances. We found
a 30-day mortality rate of 8.4% in women and 8.5% in men. Based on the data
from the National Registry of Myocardial Infarction 2 (collected between 1994
and 1998), Vaccarino et al18 reported an in-hospital
mortality rate of 16.7% and 11.5% among women and men, respectively. There
are some important differences in patient characteristics between those study
cohorts and ours. The patients included in the current study were slightly
younger but more likely to present with cardiogenic shock.18
A major difference was seen in the frequency and type of reperfusion treatment:
less than 25% of the patients in the study by Vaccarino et al18
received any reperfusion treatment, and most of those received thrombolysis.
In a recent study by Gottlieb et al,19
which included patients treated from 1992 to 1996, 30-day mortality rates
were 17.6% among women and 9.6% among men. The patients in the latter study
appear to be comparable with the patients in our study regarding age and other
entry characteristics, yet reperfusion was attempted in only about 50% of
the cases, mostly with thrombolysis.19 A higher
mortality rate for women with AMI has also been observed in population-based
studies.34 All of these reports11,14,15,18,19
indicate that patients with AMI are at higher risk if no reperfusion therapy
is applied. It is possible that women with AMI, who usually are older than
men, might suffer greater disadvantages if no reperfusion treatment is provided.
Although the life-saving role of thrombolysis as the most widely used option
for reperfusion is well established,35 many
patients are not eligible for thrombolysis,17
and these patients are at particularly high risk.36
Therefore, the interventional strategy applied among patients who did not
meet conventional eligibility criteria for thrombolysis may have contributed
to the reduction of the excess risk carried by women. In addition, the large
use of glycoprotein IIb/IIIa inhibition has probably had a positive impact
on the results of our study. Cho et al37 found
abciximab to be particularly beneficial in women undergoing coronary stenting.
An excess mortality risk among women has also been evidenced in secondary
analyses of thrombolysis trials.4- 6
A comparison of absolute mortality rates observed in those trials and in the
present study may be misleading because the thrombolysis trials generally
included younger and hemodynamically stable patients. Also, while the thrombolysis
trials have usually enrolled only patients with ST-segment elevation MI who
presented early (<6-12 hours) after symptom onset, our study encompassed
both patients with ST-segment elevation MI and those with non-ST-segment elevation
MI, patients who presented early and those who presented after 12 hours from
symptom onset. We used thrombolysis in a limited number of selected patients,
which prevents us from drawing conclusions in this regard.
After 1 year, we could not detect any significant difference in mortality
between women and men with AMI. In comparison with previous analog studies
in patients with AMI,16,19 the
use of medications such as β-blockers, angiotensin-converting enzyme
inhibitors, and statins was much more frequent in the present study. Based
on previous evidence,38- 40
the intensive pharmacological therapy has probably influenced the overall
results in this consecutive series, but we do not know whether women have
had a greater benefit than men from this regimen.
Several limitations should be acknowledged before trying to draw conclusions
from this study. First, it is a single-center study and the data presented
need replication from other centers. Second, the relatively small sample size
reduces the power of subset analyses and precluded the investigation of potentially
important interactions. Third, these results may not be generalizable to centers
that do not aggressively pursue a routine rapid reperfusion strategy mainly
based on PCI. Finally, there were too few patients treated to determine whether
similar results might have been seen had thrombolysis been the predominant
We found that women with AMI treated with a reperfusion strategy largely
based on PCI show a similar outcome with men, despite women's older age and
more adverse risk profiles. Our results suggest that sex alone should not
be a factor in deciding whether to perform primary PCI.