- Email: [email protected]
Mortality Risk and Patterns of Practice in 2,070 Patients With Acute Myocardial Infarction, 1987-92
Mortality Risk and Patterns of Practice in 2,070 Patients With Acute Myocardial Infarction, 1987-92 Relative Importance of Age, Sex, and Medical Therapy Ross T . Tsuyuki, Pharm.D. ; Koon K. Teo, M .D., Ph.D., Roland M. lkuta, M.D ., Kyung S. Bay, PhD , Paul V. Greenwood, M .D., F.C .C.P., and Terrence]. Montague , M.D . Objective: To define contemporary age- and sex-related mortality risks and patterns of medical practice in acute myocardial infarction (AMI). Design: Retrospective comparison of demographic and clinical variables, including the use of proven effective AMI medical therapy, among AMI patient cohorts from 1987 to 1992. Patients/Setting: Of a total of 2,070 AMI patients, 629 were women and 1,441, men; 951 patients were managed in university hospitals, 641 in a regional hospital, and 478 in community hospitals. Interventions: No direct study interventions; results of practice patterns and risk analyses of the earlier (198790) AMI cohorts, however, were published concurrently with the actual practices of the more recent (1991-92) cohorts and may have had some indirect effect on the recent practice patterns. Results: Univariate analysis showed that mortality was higher (p
I
n previous analyses of outcomes and patterns of practice among patients with acute myocardial infarction (AMI), we have consistently found age of 70 years or greater to be a clinically important, and statistically independent, indicator of increased relative risk for in-hospital rnortality.l'f These results confirmed similar findings from many earlier studies that have led , overall, to widespread acceptance of a direct relation between age and AMI mortality risk.4•6 Female sex has also been reported to be associated -From the Epidemiology Coordinating and Research (EPICORE) Centre. Division of Cardiology. Department of Medicine , (Drs. Teo, Ikuta, Greenwood and Montague) and the Department of Health Services Administration and Community Medicine (Dr. Bay), University of Alberta ; and the Faculty of Pharmaceutical Sciences (Dr. Tsuyuki), University of British Columbia. Vancouver, Canada. Manuscript received July 27, 1993; revision accepted November 1. Reprint requests: Dr. Montague , 2C2 Mackenzie Centre . University 0/ Alberta Hospitals. Edmonton . Alberta, Canada T6G 287
Multivariate analysis of the entire patient sample revealed age of 75 years or older (154 percent) and age 70 to 74 years (141 percent) to be associated with the highest relative risk of death in hospital. The increased relative risk associated with previous AMI was 45 percent . Acetylsalicylic acid use was associated with the greatest decrease in relative risk of death (-69 percent), followed by beta blockers (-36 percent) and thrombolysis (-31 percent). These patterns of relative risk were the same for men and women. Conclusions: Among contemporary AMI patients, advanced age and female sex are associated with relative under-utilization of proven effective medical therapy and increased risk of dying in the hospital. Although the contribution of age to AMI risk appears greater than that of gender, survival in any high risk group would likely be improved by increased use of proven medical therapy. (Chest 1994; 105:1687-92)
I
AMI=acute myocardial infarction
I
with increased early mortality risk in AMl. i . 14 In contrast to the acceptance of the age/risk relation in AMI, the relationship between femaleness and AMI risk is controversial and has not been universally accepted.9 •I 1.1S Post hoc analysis of the data from our 3 previously reported AMI study cohorts.l'f with a total of more than 900 patients, did not reveal either sex to be independently associated with increased relative mortality risk. We considered, however, that our sample size might not have been sufficiently large to reliably detect a small, but nonetheless important, sex-based mortality effect. Thus, we felt the relative contribution of sex to AMI risk remained incompletely defined. The primary purpose of this present analysis of an expanded contemporary AMI patient population was to further define the importance of gender, particularly female gender, in relation to age and to the use of proven effective medical therapy, as a risk factor for short-term mortality. CHEST/105/6/ JUNE. 1994
1687
Table
Sex Men Women Age 2:70 years <70 years Previous AMI Acetylsalicylic acid Beta blockers Thrombolysis Mortality
i-Percent Diatribution of Demographic and Clinical Variable. in N=2,070 Pallen" With Acute Myocardial Infarction by Time and Setting-
VGH 1987 (n=94)
VAH 1987 (n=1l3)
VAH 1988-89 (n"402)
VAH 1989-90 (n=342)
MIS 1991 (n=182)
RCH 1990-91 (n-336)
RCH 1991-92 (n'"'305)
SMH 1991-92 (n-296)
Mean
71 29
65 35
74 26
73 27
69 31
66 34
66 34
69 31
70 30
46 54 31 55
51 49 41 34
33 67 27 71
39 61 29 83
37 63 25 92
44 57 30 79
50 50 33
39 61 27
80
77
41 59 30 76
64 17 20
27 10 20
41 15 14
57 34 13
34 51 12
35 24 10
40 30 18
56 30 14
44 27 14
·MIS= Misericordia Hospital . Edmonton; RCH=Royal Columbian Hospital. New Westminster; SMHIIIMemorial Hospital. Surrey ; VAH=University of Alberta Hospitals. Edmonton; and VGH=Victoria General Hospital . HalIfax. All in Canada.
PATIENTS AND METHODS
The study population consisted of 2.070 patients with a clinical diagnosis of AMI at discharge or death between 1987 and 1992 (Table 1). The total study population was composed of four AMI cohorts previously reported from the University of Alberta Hospitals. Edmonton. Canada and the Victoria General Hospital. Halifax. Canada. between 1987 and 1990 (n=951),1-3 plus four new AMI patient cohorts recruited in 1991 and 1992 (n=I.119) (Table 1). The new AMI cohorts included two consecutive series of AMI patients from the Royal Columbian Hospital. a regional hospital in New Westminster. Canada. Patients in the first of these two series had AMI in a 12-month period from 1 September 1990 to 31 August 1991 (n=336) and in the second had AMI between 1 September 1991 and 1 June 1992 (n=305) (Table 1). The third new cohort was composed of 296 consecutive patients with AMI between 1 June 1991 and 1 June 1992 at the Surrey Memorial Hospital. a community hospital in Surrey. Canada (Table 1). The fourth new AMI cohort (n=182) was diagnosed and managed at the Misericordia Hospital. a community hospital in Edmonton. Canada between 1 January 1991 and 31 December 1991 (Table 1). The total AMI patient sample had 629 women and 1,441 men (Table 1). The average age of the women was 71 years; male patients averaged 63 years of age. Twelve hundred twenty-four AMI patients were less than 70 years of age and 846 patients were 70 years of age or older. The criteria for the diagnosis of AMI were common to all study centers and required the presence of at least two of the follow ing cardiac-compatible visceral pain, diagnostic electrocardiographic findings and elevations of serum levels of cardiac enzymes. l-3 The principal outcome variables have been previously described.·-3 For the purposes of this study they. however, were limited to measures available in all cohorts and settings. The variables were age. sex, history of previous AMI, in-hospital use of proven effective medical therapy (thrombolysis. beta blockers. acetylsalicylic acid). and in-hospital survival/mortality. Effective medical therapy was defined as medication proven to reduce mortality in large randomized, controlled AMI clinical trials before 1993 . Inferential univariate comparisons of continuous data between the female and male study patients were done using two-tailed unpaired Student's t tests ; nominal scale variables were compared 1888
with chi-square tests . For the total AMI patient sample. as well as for the female and male subgroups and the two age-based subgroups (70 years and older vs younger than 70 years of age) . the association between relative mortality risk, as the dependent variable, and all other variables. as independent variables, was assessed by multiple logistic regression with discriminant analy-
ses.
RESULTS
While there were small quantitative differences, the qualitative distribution of medication use and mortality data were remarkably similar across the geographic and temporal spectra of the study. In light of this marked inter-eohort similarity, the data from the various cohorts were combined to allow more reliable assessment of the relation between mortality and the other study variables, in the patient group as a whole and within the subgroups of older and younger patients and men and women. The distribution of demographic and clinical variables in the various AMI patient cohorts is shown in Table 1. Overall, 70 percent of the patients were men and 41 percent were 70 years of age or older (Table 1).The female patient group averaged 71 years of age and the male group, 63 years. Previous AMI had occurred, on average, in 30 percent of the patients (Table 1). Acetylsalicylic acid was the most widely used of the proven AMI therapies, averaging 76 percent, with a range from 34 percent in 1987 to 92 percent in 1991 (Table 1). Beta blocker use ranged from a low of 27 percent in 1987 to highs of 57 percent in 1989-90and 56 percent in 1992,averaging 44 percent overall. Thrombolysis was used, on average, in 27 percent of all patients, with a range of 10 percent (1989) to 51 percent (1991) (Table 1). Overall, in-hospital mortality averaged 14 percent, with a temporal decrement from 20 percent in the Reviewof 2.070 PatIen1s WIthAaJl8 MyocardIal Infarction (TsuyrJd et aI)
Table 3-Univariate Comparilon of Demographic and Clinical Variable. in 846 Patient. 70 Year. of Age and Older and 1,224 Patien" Younger Than 70 Yeara With Acute Myocardial Infarction, 1987 to 1992
Table 2-Univariate Comparilon of Demographic and Clinical Variable. in 629 Women and 1,441 Men With Acute MyocardiIJllnfarction, 1987 to 1992 Women,9Ii
p Value
Men,9Ii
~70
Age ~70
<70 Previous AMI Acetylsalicylic acid Beta blockers Thrombolysis Mortality
61 39 27 69 36 20
18
32 68 31 79 48 30 12
Years, %
<70 Years, %
p Value
80 20 34 83 52 34 7
<0 .0001 <0.0001 <0 .0001 <0 .0001 <0 .0001 <0 .0001 <0 .0001
Sex
<0.0001 <0 .0001 NS* <0 .0001 <0 .0001 <0 .0001 <0 .0001
Men Women Previous AMI Acetylsalicylic acid Beta blockers Thrombolysis Mortality
*NS=not significant.
1987 cohorts to a low of 10 percent in a 1990-91 cohort (Table 1). The AMI survivors to hospital discharge averaged 64 years of age, vs 73 years for nonsurvivors. The univariate comparisons of demographic and clinical variables between AMI men and women, and between AMI patients younger than 70 years of age and 70 years and older, are shown in Tables 2 and 3, respectively. In summary, the data showed that AMI women were older than AMI men, or conversely, that the older AMI patient group had a much higher proportion of women (Tables 2 and 3). Both AMI women (Table 2) and AMI patients 70 years of age and older (Table 3) had significantly less prescription of all medications previously proven to be effective in reducing AMI mortality. Mortality was significantly higher in AMI women, relative to AMI men (Table 2), and in older, relative to younger (Table 3), AMI patients. The degree of difference in mortality/survival, however, was much
55 45 26 65 33 16 25
greater between the two age subgroups, compared with the two gender subgroups (Tables 2 and 3). Multivariate analysis of the total AMI population of 2,070 patients revealed age of 75 years or older, age between 70 and 74 years, history of previous AMI, and age between 65 and 69 years to be independently associated with increased relative risk of death (Fig 1). The use of acetylsalicylic acid, beta blockers, and thrombolysis, on the other hand were associated with decreased relative mortality risk (Fig 1). Patient sex was not independently associated with AMI mortality (Fig 1). As indicated in Figure 2, the same pattern of agebased relative mortality risk was present in both women and men AMI subjects. When age was removed as an independent variable (Fig 3), multiple logistic regression analyses continued to show the same relative distribution of mortality risk for the remaining variables. In particular, use of acetylsalicylic acid was associated with the lowest relative
RELATIVE RISK· ALL SUBJECTS ("=2070) 100 , - - - - - - - - - - - - - - - - - - - - - - - - - ,
* p<0.05
*
150 1l7"TJ"TT7TrT1
100
%
1. Multiple logistic regression analysis of demographic and clinical variables, relative to mortality risk in the total AMI study population of 2,070 patienls from 1987 to 1992. ASA=acetylsalicylic acid ; Bblkr=beta blocker; Iysis=thrombolysis; prev= previous. FIGURE
50
.50
* AGE 75+
PREY AMI
AGE 60·69
LYSIS
BBlkr
CHEST /105/6/ JUNE, 1994
ASA
1889
RELATIVE RISK· AGE 70+ (n-848)
..r-----------------------, * p
... ...--------------------., RELATIVE RISK - FEMALES (n=82e)
* p
*
••
" ...
.. ....
p"ev
..s..
Lyel.
AMI
... , . - - - - - - - - - - - - - - - - - - - - - - - , RELATIVE RISK· AGE <70 (n"'1224)
..
RELATIVE RISK - MALES (n=1441)
...----------------------, *
* p
*
p
. "
••
•
... ...
.-
* ~"II!Y
AMI
...... .. Oil!
LYaI•
.-
.,. AU
2. Multiple logistic regression anal ysis of demographic and clinical variables, relative to mortality, in female (top panel) and male (bottom panel) AMI stud y subjects. ASA=acel ylsalicylic acid ; Bblkr=beta blocker ; Iysis=thrombolysis; pre v=previous. FIGURE
mortality risk in all patient subgroups, averaging minus 70 percent (Fig 1 to 3). DISCUSSION
The results of the present analysis of patterns of medical practice and patient outcomes in a total of 2,070 AMI patients confirmed and extended the findings from our three previous studies. I . 3 Specifically, age was an independent risk factor in both male and female AMI patients, and older patients did not receive as much proven medical therapy as younger patients, despite their being at significantly higher risk. Women also had a higher mortality than men and received significantly less of all proven medical therapies. Moreover, these age- and sex-related patterns of AMI risk and medical practice were very similar across the spectra of clinical settings and time embraced by the study. Representative results from previous studies comparing in-hospital male and female risk in AMI patients from 1953 to 1985 are summarized in Table 4.7• 12 Univariate analyses revealed higher in-hospital 1890
aeM:
,0.,.
ACIa
7..
~ ... y AMI
AeM:
....
Ln..
....,
AU
F IGURE 3. Multiple logistic regression analysis of demographic and clinical variables, relati ve to mortality, in AMI patients 70 years of age and greater (top panel) and less than 70 years (botlorn panel) . ASA=acelylsalic ylic acid; Bblkr=bela blocker ; lysis =lhrombolysis; prev = previous.
mortality for women, compared with men, in all six studies (Table 4). Only two studies,9,10 however, suggested that female sex was independent of other known risk factors and in one these studies, with a sample size of only 816 patients, the results may have been the result of chance.? In a follow-up editorial, the investigators suggested that the apparent higher female AMI mortality was primarily related to being black , rather than female gender.P The sample size and results of the present AMI anal ysis were very similar to those of the study of Dittrich et al" (Table 4). Specifically, female sex, although associated with increased risk shown by univariate analysis, was not associated with increased relative risk by multivariate analysis. In general, the results of the previous studies with smaller sample sizes7,8,12 are also concordant with our findings and those of Dittrich et al. 11 The largest reported study sample comprised 5,839 AMI patients from Israel with AMI during 1981 to 1983.10 The Israeli study, with nearly three times the sample size, and more than three times the average Review of 2.070 Patients Wrtll Acute Myocardial Infarction (Tsuyuki 8t aI)
Table 4-Summary of Previous Studies of Effect of Gender on Prognosis in Acute Myocardial Infarction Univariate Comparison of In-hospital Mortality Study Hughes?
Fiebach" Tofler?
Greenland'? Dittrich'! Robinson 12_
Study Interval
No. of Patients
Women, %
Men, %
p value
1953-60 1978-82 1978-83 1981-83 1979-84 1980-85
445 1122 816 5839 2089 980
40 14 13 23 18 13
31 9 7 16 12 7
0.01 0.05 0.005 0.003 0.002
Multivariate Analysis of Femal e Gender as Independent Risk Factor 0 0
+ + 0 0
-Included patients with unstable angina and acute myocardial infarction.
event rate, of any other study should have the greatest statistical power and least chance of beta error.l? In contrast to the majority of the smaller studies , the Israeli investigators found that female sex was an independent contributor to AMI mortality risk (Table 4).10
Although at first glance the findings of the various AMI risk analyses might appear discordant, they are, in fact , compatible. In particular, the dichotomy between univariate and multivariate analyses of risk for female sex in AMI is probably more apparent than real. The dichotomy arises in large measure because the statistic tools of multivariate analyses are often interpreted as truly discriminating, in a qualitative fashion, pathophysiologic contributions to risk. A more correct interpretation is that multivariate analysis discriminates hierarchical, but still quantitative, contributions to risk. Unlike female sex, advanced age in AMI populations has been repeatedly found to be "independent" of other baseline variables in multivariate analyses only because it makes a quantitatively greater contribution to mortality risk than does female sex. The distinct likelihood is that if the sample sizes in all the studies were as large, and the mortality rates as high, as in the study by Greenland et al,1O multivariate analysis would have showen female sex to be associated with relatively increased AMI mortality risk. The previous studies on gender-related AMI risk summarized in Table 4 entered patients from 10 to 40 years ago-before the delineation of trials-proven modern AMI therapy.l'' The question might well be asked, were there also age- and sex-related differences in application of therapy, albeit therapy by consensus, during the time frame of those previous AMI studies? If so, did they contribute to outcomes? It is, of course, impossible to be absolutely certain about either answer, but it is unlikely that any sex or age-based differences in the AMI therapy in past decades caused any significant differences in outcomes. As Radzik and his colleagues'? from the Montreal Heart Institute have recently pointed out, there was essentially no systematic application of
thrombolysis, intravenous nitroglycerin, heparin, or acetylsalicylic acid therapy a decade ago, even in leading cardiac tertiary care centers. Now that large scale clinical trials, however, have shown several therapies to significantly reduce death risk in AMI, would it equalize contemporary older/ younger and female/male mortality risks if age- and sex-related therapeutic differences were reduced or eliminated? Although it is somewhat of a circular argument, the answer would appear to be no. From review of the literature, women 7- 12 (Table 4) and older patients 4-6 had significantly higher AMI mortality risks before the development of thrombolysis and the other proven, current AMI medical therapies. On the other hand, the global and subgroup analyses of contemporary AMI medical practice in the present study were very consistent-use of previously proven medical therapy was associated with reduced risk in all data subsets (Fig 1-3). It is also reasonable to assume that proven efficacious agents will usually have their greatest effect in the highest risk groups, where there is the greatest pathophysiologic disturbance and , concomitantly, the greatest opportunity for positive modification by the effective agent. 3,6 If this line of reasoning is correct, then enhanced use of effective medical therapy in any high risk group, such as the elderly or women, should be associated with decreased mortality risk. To reliably test this hypothesis, further studies are needed. REFERENCES
1 Montague T . Wong R. Crowell R. Bay K. Marshall 0, Tymchak W, et al. Acute myocardial infarction: contemporary risk and management in older versus younger patients. Can J Cardiol 1990; 6:241-46 2 Montague TJ, Ikuta RM, Wong RY. Bay KS. Teo KK. Davies NJ. Comparison of risk and patterns of practice in patients older and younger than 70 years with acute myocardial infarction in a 2-year period (1987-1989). Am J Cardiol 1991; 68:843-47 3 Montague TJ, Wong RY, Burton JR, Bay KS, Catellier OJ. Teo KK. Changes in acute myocardial infarction risk and patterns of practice for patients older and younger than 70 years, 1987CHEST/105/6/ JUNE, 1994
1891
90. Can 1 Card ioll992; 8:596-600 4 Cairns lA, Singer 1, Gent M, Holder DA, Rogers D, Sackett DL, et al. Coronary care unit utilization in Hamilton, Ontario, a city of 375,000 people. Can 1 Cardiol 1988; 4:25-32 5 Smith SC, Gilpin E, Ahnve S, Dittrich H, Nicod P, Henning H, et al. Outlook after myocardial infarction in the very elderly compared with that in patients aged 65 to 75 years. 1 Am Coil Cardiol 1990; 16:784-92 6 ISI5-2 (Second International Study of Infarct Survival) Collaborative Group . Randomized trial of intravenous streptok inase, oral aspirin, both or neither among 17,187 cases of suspected acute myocardial infarction. Lancet 1988; ii:349-60 7 Hughes WL , Kalblleisch 1M, Brandt EN, Costiloe IP. Myocardial infarction prognosis by discriminant analysis. Arch Int Med 1963; 111:120-27 8 Fiebach NH , Viscoli CM, Horowitz RI. Differences between women and men in survival after myocardial infarction : biology or methodology? lAMA 1990; 263:1092-96 9 Toller GH, Stone PH , Muller IE, the MILlS study group . Effects of gender and race on prognosis after myocardial infarction: adverse prognosis for women, particularly black women. 1 Am Coli Cardiol1987; 9:473-82 10 Greenland P, Reicher-Reiss H, Goldbourt U, Behar S, the Israeli SPRINT Investigators. In-hospital and l-year mortality in 1,524 women after myocardial infarction: comparison with 4,315
men. Circulation 1991; 83:484-91 11 Dittrich H, Gilpin E, Nicod P, Cali G, Henning H, Ross1. Acute myocardial infarction in women : influence of gender on mortality and prognostic variables. Am 1 CardioI1988; 62:1-7 12 Robinson K, Conroy RM, Mulcachy R, Hickey N. Risk factors and in-hospital course of first episode of myocardial infarction or coronary insufficiency in women. 1 Am Coli CardioI1988; 11:932-36 13 Puletti M, Sunseri L, Curione M, Erba SM, Borgia C. Acute myocardial infarction: sex-related differences in prognosis. Am Heart 1 1984; 108:63-6 14 Schor S, Shani M, Modan B. Factors affecting immediate mortality of patients with acute myocardial infarction:a nationwide study. Chest 1975; 68:217-21 15 Toller GH , Stone PH , Muller.IE, Braunwald E (MILlS Study Group). Mortality for women after acute myocardial infarction . Am I Cardio11989; 64:256 16 Yusuf S, Wittes 1, Friedman L. Overview of results of randomized clinical trials in heart disease: treatments following myocardial infarction. lAMA 1988; 260:2088-93 17 Radzik D, Talajic M, Roy D, Hii ITY, Thibault B, Theroux P, et aI. Have changes in the management of acute myocardial infarction between 1980 and 1990 reduced late arrhythmic complications [abstract]? 1 Am Coli Cardiol 1993; 21(suppl A):43A
Thrombosis of Prosthetic Valves The George Gabor Foundation will present this conference in Budapest, Hungary, September 15-17 at the Hotel Thermal Helia. For information, contact Natyas Jektaum MD, Hungarian Institute of Cardiology, PO Box 88, H-1450 Budapest, Hungary.
1892
Reviewof 2,070 Patients WIth Acuts MyocardIaIlnIarctIon (TsuyuId at B1)
I
n previous analyses of outcomes and patterns of practice among patients with acute myocardial infarction (AMI), we have consistently found age of 70 years or greater to be a clinically important, and statistically independent, indicator of increased relative risk for in-hospital rnortality.l'f These results confirmed similar findings from many earlier studies that have led , overall, to widespread acceptance of a direct relation between age and AMI mortality risk.4•6 Female sex has also been reported to be associated -From the Epidemiology Coordinating and Research (EPICORE) Centre. Division of Cardiology. Department of Medicine , (Drs. Teo, Ikuta, Greenwood and Montague) and the Department of Health Services Administration and Community Medicine (Dr. Bay), University of Alberta ; and the Faculty of Pharmaceutical Sciences (Dr. Tsuyuki), University of British Columbia. Vancouver, Canada. Manuscript received July 27, 1993; revision accepted November 1. Reprint requests: Dr. Montague , 2C2 Mackenzie Centre . University 0/ Alberta Hospitals. Edmonton . Alberta, Canada T6G 287
Multivariate analysis of the entire patient sample revealed age of 75 years or older (154 percent) and age 70 to 74 years (141 percent) to be associated with the highest relative risk of death in hospital. The increased relative risk associated with previous AMI was 45 percent . Acetylsalicylic acid use was associated with the greatest decrease in relative risk of death (-69 percent), followed by beta blockers (-36 percent) and thrombolysis (-31 percent). These patterns of relative risk were the same for men and women. Conclusions: Among contemporary AMI patients, advanced age and female sex are associated with relative under-utilization of proven effective medical therapy and increased risk of dying in the hospital. Although the contribution of age to AMI risk appears greater than that of gender, survival in any high risk group would likely be improved by increased use of proven medical therapy. (Chest 1994; 105:1687-92)
I
AMI=acute myocardial infarction
I
with increased early mortality risk in AMl. i . 14 In contrast to the acceptance of the age/risk relation in AMI, the relationship between femaleness and AMI risk is controversial and has not been universally accepted.9 •I 1.1S Post hoc analysis of the data from our 3 previously reported AMI study cohorts.l'f with a total of more than 900 patients, did not reveal either sex to be independently associated with increased relative mortality risk. We considered, however, that our sample size might not have been sufficiently large to reliably detect a small, but nonetheless important, sex-based mortality effect. Thus, we felt the relative contribution of sex to AMI risk remained incompletely defined. The primary purpose of this present analysis of an expanded contemporary AMI patient population was to further define the importance of gender, particularly female gender, in relation to age and to the use of proven effective medical therapy, as a risk factor for short-term mortality. CHEST/105/6/ JUNE. 1994
1687
Table
Sex Men Women Age 2:70 years <70 years Previous AMI Acetylsalicylic acid Beta blockers Thrombolysis Mortality
i-Percent Diatribution of Demographic and Clinical Variable. in N=2,070 Pallen" With Acute Myocardial Infarction by Time and Setting-
VGH 1987 (n=94)
VAH 1987 (n=1l3)
VAH 1988-89 (n"402)
VAH 1989-90 (n=342)
MIS 1991 (n=182)
RCH 1990-91 (n-336)
RCH 1991-92 (n'"'305)
SMH 1991-92 (n-296)
Mean
71 29
65 35
74 26
73 27
69 31
66 34
66 34
69 31
70 30
46 54 31 55
51 49 41 34
33 67 27 71
39 61 29 83
37 63 25 92
44 57 30 79
50 50 33
39 61 27
80
77
41 59 30 76
64 17 20
27 10 20
41 15 14
57 34 13
34 51 12
35 24 10
40 30 18
56 30 14
44 27 14
·MIS= Misericordia Hospital . Edmonton; RCH=Royal Columbian Hospital. New Westminster; SMHIIIMemorial Hospital. Surrey ; VAH=University of Alberta Hospitals. Edmonton; and VGH=Victoria General Hospital . HalIfax. All in Canada.
PATIENTS AND METHODS
The study population consisted of 2.070 patients with a clinical diagnosis of AMI at discharge or death between 1987 and 1992 (Table 1). The total study population was composed of four AMI cohorts previously reported from the University of Alberta Hospitals. Edmonton. Canada and the Victoria General Hospital. Halifax. Canada. between 1987 and 1990 (n=951),1-3 plus four new AMI patient cohorts recruited in 1991 and 1992 (n=I.119) (Table 1). The new AMI cohorts included two consecutive series of AMI patients from the Royal Columbian Hospital. a regional hospital in New Westminster. Canada. Patients in the first of these two series had AMI in a 12-month period from 1 September 1990 to 31 August 1991 (n=336) and in the second had AMI between 1 September 1991 and 1 June 1992 (n=305) (Table 1). The third new cohort was composed of 296 consecutive patients with AMI between 1 June 1991 and 1 June 1992 at the Surrey Memorial Hospital. a community hospital in Surrey. Canada (Table 1). The fourth new AMI cohort (n=182) was diagnosed and managed at the Misericordia Hospital. a community hospital in Edmonton. Canada between 1 January 1991 and 31 December 1991 (Table 1). The total AMI patient sample had 629 women and 1,441 men (Table 1). The average age of the women was 71 years; male patients averaged 63 years of age. Twelve hundred twenty-four AMI patients were less than 70 years of age and 846 patients were 70 years of age or older. The criteria for the diagnosis of AMI were common to all study centers and required the presence of at least two of the follow ing cardiac-compatible visceral pain, diagnostic electrocardiographic findings and elevations of serum levels of cardiac enzymes. l-3 The principal outcome variables have been previously described.·-3 For the purposes of this study they. however, were limited to measures available in all cohorts and settings. The variables were age. sex, history of previous AMI, in-hospital use of proven effective medical therapy (thrombolysis. beta blockers. acetylsalicylic acid). and in-hospital survival/mortality. Effective medical therapy was defined as medication proven to reduce mortality in large randomized, controlled AMI clinical trials before 1993 . Inferential univariate comparisons of continuous data between the female and male study patients were done using two-tailed unpaired Student's t tests ; nominal scale variables were compared 1888
with chi-square tests . For the total AMI patient sample. as well as for the female and male subgroups and the two age-based subgroups (70 years and older vs younger than 70 years of age) . the association between relative mortality risk, as the dependent variable, and all other variables. as independent variables, was assessed by multiple logistic regression with discriminant analy-
ses.
RESULTS
While there were small quantitative differences, the qualitative distribution of medication use and mortality data were remarkably similar across the geographic and temporal spectra of the study. In light of this marked inter-eohort similarity, the data from the various cohorts were combined to allow more reliable assessment of the relation between mortality and the other study variables, in the patient group as a whole and within the subgroups of older and younger patients and men and women. The distribution of demographic and clinical variables in the various AMI patient cohorts is shown in Table 1. Overall, 70 percent of the patients were men and 41 percent were 70 years of age or older (Table 1).The female patient group averaged 71 years of age and the male group, 63 years. Previous AMI had occurred, on average, in 30 percent of the patients (Table 1). Acetylsalicylic acid was the most widely used of the proven AMI therapies, averaging 76 percent, with a range from 34 percent in 1987 to 92 percent in 1991 (Table 1). Beta blocker use ranged from a low of 27 percent in 1987 to highs of 57 percent in 1989-90and 56 percent in 1992,averaging 44 percent overall. Thrombolysis was used, on average, in 27 percent of all patients, with a range of 10 percent (1989) to 51 percent (1991) (Table 1). Overall, in-hospital mortality averaged 14 percent, with a temporal decrement from 20 percent in the Reviewof 2.070 PatIen1s WIthAaJl8 MyocardIal Infarction (TsuyrJd et aI)
Table 3-Univariate Comparilon of Demographic and Clinical Variable. in 846 Patient. 70 Year. of Age and Older and 1,224 Patien" Younger Than 70 Yeara With Acute Myocardial Infarction, 1987 to 1992
Table 2-Univariate Comparilon of Demographic and Clinical Variable. in 629 Women and 1,441 Men With Acute MyocardiIJllnfarction, 1987 to 1992 Women,9Ii
p Value
Men,9Ii
~70
Age ~70
<70 Previous AMI Acetylsalicylic acid Beta blockers Thrombolysis Mortality
61 39 27 69 36 20
18
32 68 31 79 48 30 12
Years, %
<70 Years, %
p Value
80 20 34 83 52 34 7
<0 .0001 <0.0001 <0 .0001 <0 .0001 <0 .0001 <0 .0001 <0 .0001
Sex
<0.0001 <0 .0001 NS* <0 .0001 <0 .0001 <0 .0001 <0 .0001
Men Women Previous AMI Acetylsalicylic acid Beta blockers Thrombolysis Mortality
*NS=not significant.
1987 cohorts to a low of 10 percent in a 1990-91 cohort (Table 1). The AMI survivors to hospital discharge averaged 64 years of age, vs 73 years for nonsurvivors. The univariate comparisons of demographic and clinical variables between AMI men and women, and between AMI patients younger than 70 years of age and 70 years and older, are shown in Tables 2 and 3, respectively. In summary, the data showed that AMI women were older than AMI men, or conversely, that the older AMI patient group had a much higher proportion of women (Tables 2 and 3). Both AMI women (Table 2) and AMI patients 70 years of age and older (Table 3) had significantly less prescription of all medications previously proven to be effective in reducing AMI mortality. Mortality was significantly higher in AMI women, relative to AMI men (Table 2), and in older, relative to younger (Table 3), AMI patients. The degree of difference in mortality/survival, however, was much
55 45 26 65 33 16 25
greater between the two age subgroups, compared with the two gender subgroups (Tables 2 and 3). Multivariate analysis of the total AMI population of 2,070 patients revealed age of 75 years or older, age between 70 and 74 years, history of previous AMI, and age between 65 and 69 years to be independently associated with increased relative risk of death (Fig 1). The use of acetylsalicylic acid, beta blockers, and thrombolysis, on the other hand were associated with decreased relative mortality risk (Fig 1). Patient sex was not independently associated with AMI mortality (Fig 1). As indicated in Figure 2, the same pattern of agebased relative mortality risk was present in both women and men AMI subjects. When age was removed as an independent variable (Fig 3), multiple logistic regression analyses continued to show the same relative distribution of mortality risk for the remaining variables. In particular, use of acetylsalicylic acid was associated with the lowest relative
RELATIVE RISK· ALL SUBJECTS ("=2070) 100 , - - - - - - - - - - - - - - - - - - - - - - - - - ,
* p<0.05
*
150 1l7"TJ"TT7TrT1
100
%
1. Multiple logistic regression analysis of demographic and clinical variables, relative to mortality risk in the total AMI study population of 2,070 patienls from 1987 to 1992. ASA=acetylsalicylic acid ; Bblkr=beta blocker; Iysis=thrombolysis; prev= previous. FIGURE
50
.50
* AGE 75+
PREY AMI
AGE 60·69
LYSIS
BBlkr
CHEST /105/6/ JUNE, 1994
ASA
1889
RELATIVE RISK· AGE 70+ (n-848)
..r-----------------------, * p
... ...--------------------., RELATIVE RISK - FEMALES (n=82e)
* p
*
••
" ...
.. ....
p"ev
..s..
Lyel.
AMI
... , . - - - - - - - - - - - - - - - - - - - - - - - , RELATIVE RISK· AGE <70 (n"'1224)
..
RELATIVE RISK - MALES (n=1441)
...----------------------, *
* p
*
p
. "
••
•
... ...
.-
* ~"II!Y
AMI
...... .. Oil!
LYaI•
.-
.,. AU
2. Multiple logistic regression anal ysis of demographic and clinical variables, relative to mortality, in female (top panel) and male (bottom panel) AMI stud y subjects. ASA=acel ylsalicylic acid ; Bblkr=beta blocker ; Iysis=thrombolysis; pre v=previous. FIGURE
mortality risk in all patient subgroups, averaging minus 70 percent (Fig 1 to 3). DISCUSSION
The results of the present analysis of patterns of medical practice and patient outcomes in a total of 2,070 AMI patients confirmed and extended the findings from our three previous studies. I . 3 Specifically, age was an independent risk factor in both male and female AMI patients, and older patients did not receive as much proven medical therapy as younger patients, despite their being at significantly higher risk. Women also had a higher mortality than men and received significantly less of all proven medical therapies. Moreover, these age- and sex-related patterns of AMI risk and medical practice were very similar across the spectra of clinical settings and time embraced by the study. Representative results from previous studies comparing in-hospital male and female risk in AMI patients from 1953 to 1985 are summarized in Table 4.7• 12 Univariate analyses revealed higher in-hospital 1890
aeM:
,0.,.
ACIa
7..
~ ... y AMI
AeM:
....
Ln..
....,
AU
F IGURE 3. Multiple logistic regression analysis of demographic and clinical variables, relati ve to mortality, in AMI patients 70 years of age and greater (top panel) and less than 70 years (botlorn panel) . ASA=acelylsalic ylic acid; Bblkr=bela blocker ; lysis =lhrombolysis; prev = previous.
mortality for women, compared with men, in all six studies (Table 4). Only two studies,9,10 however, suggested that female sex was independent of other known risk factors and in one these studies, with a sample size of only 816 patients, the results may have been the result of chance.? In a follow-up editorial, the investigators suggested that the apparent higher female AMI mortality was primarily related to being black , rather than female gender.P The sample size and results of the present AMI anal ysis were very similar to those of the study of Dittrich et al" (Table 4). Specifically, female sex, although associated with increased risk shown by univariate analysis, was not associated with increased relative risk by multivariate analysis. In general, the results of the previous studies with smaller sample sizes7,8,12 are also concordant with our findings and those of Dittrich et al. 11 The largest reported study sample comprised 5,839 AMI patients from Israel with AMI during 1981 to 1983.10 The Israeli study, with nearly three times the sample size, and more than three times the average Review of 2.070 Patients Wrtll Acute Myocardial Infarction (Tsuyuki 8t aI)
Table 4-Summary of Previous Studies of Effect of Gender on Prognosis in Acute Myocardial Infarction Univariate Comparison of In-hospital Mortality Study Hughes?
Fiebach" Tofler?
Greenland'? Dittrich'! Robinson 12_
Study Interval
No. of Patients
Women, %
Men, %
p value
1953-60 1978-82 1978-83 1981-83 1979-84 1980-85
445 1122 816 5839 2089 980
40 14 13 23 18 13
31 9 7 16 12 7
0.01 0.05 0.005 0.003 0.002
Multivariate Analysis of Femal e Gender as Independent Risk Factor 0 0
+ + 0 0
-Included patients with unstable angina and acute myocardial infarction.
event rate, of any other study should have the greatest statistical power and least chance of beta error.l? In contrast to the majority of the smaller studies , the Israeli investigators found that female sex was an independent contributor to AMI mortality risk (Table 4).10
Although at first glance the findings of the various AMI risk analyses might appear discordant, they are, in fact , compatible. In particular, the dichotomy between univariate and multivariate analyses of risk for female sex in AMI is probably more apparent than real. The dichotomy arises in large measure because the statistic tools of multivariate analyses are often interpreted as truly discriminating, in a qualitative fashion, pathophysiologic contributions to risk. A more correct interpretation is that multivariate analysis discriminates hierarchical, but still quantitative, contributions to risk. Unlike female sex, advanced age in AMI populations has been repeatedly found to be "independent" of other baseline variables in multivariate analyses only because it makes a quantitatively greater contribution to mortality risk than does female sex. The distinct likelihood is that if the sample sizes in all the studies were as large, and the mortality rates as high, as in the study by Greenland et al,1O multivariate analysis would have showen female sex to be associated with relatively increased AMI mortality risk. The previous studies on gender-related AMI risk summarized in Table 4 entered patients from 10 to 40 years ago-before the delineation of trials-proven modern AMI therapy.l'' The question might well be asked, were there also age- and sex-related differences in application of therapy, albeit therapy by consensus, during the time frame of those previous AMI studies? If so, did they contribute to outcomes? It is, of course, impossible to be absolutely certain about either answer, but it is unlikely that any sex or age-based differences in the AMI therapy in past decades caused any significant differences in outcomes. As Radzik and his colleagues'? from the Montreal Heart Institute have recently pointed out, there was essentially no systematic application of
thrombolysis, intravenous nitroglycerin, heparin, or acetylsalicylic acid therapy a decade ago, even in leading cardiac tertiary care centers. Now that large scale clinical trials, however, have shown several therapies to significantly reduce death risk in AMI, would it equalize contemporary older/ younger and female/male mortality risks if age- and sex-related therapeutic differences were reduced or eliminated? Although it is somewhat of a circular argument, the answer would appear to be no. From review of the literature, women 7- 12 (Table 4) and older patients 4-6 had significantly higher AMI mortality risks before the development of thrombolysis and the other proven, current AMI medical therapies. On the other hand, the global and subgroup analyses of contemporary AMI medical practice in the present study were very consistent-use of previously proven medical therapy was associated with reduced risk in all data subsets (Fig 1-3). It is also reasonable to assume that proven efficacious agents will usually have their greatest effect in the highest risk groups, where there is the greatest pathophysiologic disturbance and , concomitantly, the greatest opportunity for positive modification by the effective agent. 3,6 If this line of reasoning is correct, then enhanced use of effective medical therapy in any high risk group, such as the elderly or women, should be associated with decreased mortality risk. To reliably test this hypothesis, further studies are needed. REFERENCES
1 Montague T . Wong R. Crowell R. Bay K. Marshall 0, Tymchak W, et al. Acute myocardial infarction: contemporary risk and management in older versus younger patients. Can J Cardiol 1990; 6:241-46 2 Montague TJ, Ikuta RM, Wong RY. Bay KS. Teo KK. Davies NJ. Comparison of risk and patterns of practice in patients older and younger than 70 years with acute myocardial infarction in a 2-year period (1987-1989). Am J Cardiol 1991; 68:843-47 3 Montague TJ, Wong RY, Burton JR, Bay KS, Catellier OJ. Teo KK. Changes in acute myocardial infarction risk and patterns of practice for patients older and younger than 70 years, 1987CHEST/105/6/ JUNE, 1994
1891
90. Can 1 Card ioll992; 8:596-600 4 Cairns lA, Singer 1, Gent M, Holder DA, Rogers D, Sackett DL, et al. Coronary care unit utilization in Hamilton, Ontario, a city of 375,000 people. Can 1 Cardiol 1988; 4:25-32 5 Smith SC, Gilpin E, Ahnve S, Dittrich H, Nicod P, Henning H, et al. Outlook after myocardial infarction in the very elderly compared with that in patients aged 65 to 75 years. 1 Am Coil Cardiol 1990; 16:784-92 6 ISI5-2 (Second International Study of Infarct Survival) Collaborative Group . Randomized trial of intravenous streptok inase, oral aspirin, both or neither among 17,187 cases of suspected acute myocardial infarction. Lancet 1988; ii:349-60 7 Hughes WL , Kalblleisch 1M, Brandt EN, Costiloe IP. Myocardial infarction prognosis by discriminant analysis. Arch Int Med 1963; 111:120-27 8 Fiebach NH , Viscoli CM, Horowitz RI. Differences between women and men in survival after myocardial infarction : biology or methodology? lAMA 1990; 263:1092-96 9 Toller GH, Stone PH , Muller IE, the MILlS study group . Effects of gender and race on prognosis after myocardial infarction: adverse prognosis for women, particularly black women. 1 Am Coli Cardiol1987; 9:473-82 10 Greenland P, Reicher-Reiss H, Goldbourt U, Behar S, the Israeli SPRINT Investigators. In-hospital and l-year mortality in 1,524 women after myocardial infarction: comparison with 4,315
men. Circulation 1991; 83:484-91 11 Dittrich H, Gilpin E, Nicod P, Cali G, Henning H, Ross1. Acute myocardial infarction in women : influence of gender on mortality and prognostic variables. Am 1 CardioI1988; 62:1-7 12 Robinson K, Conroy RM, Mulcachy R, Hickey N. Risk factors and in-hospital course of first episode of myocardial infarction or coronary insufficiency in women. 1 Am Coli CardioI1988; 11:932-36 13 Puletti M, Sunseri L, Curione M, Erba SM, Borgia C. Acute myocardial infarction: sex-related differences in prognosis. Am Heart 1 1984; 108:63-6 14 Schor S, Shani M, Modan B. Factors affecting immediate mortality of patients with acute myocardial infarction:a nationwide study. Chest 1975; 68:217-21 15 Toller GH , Stone PH , Muller.IE, Braunwald E (MILlS Study Group). Mortality for women after acute myocardial infarction . Am I Cardio11989; 64:256 16 Yusuf S, Wittes 1, Friedman L. Overview of results of randomized clinical trials in heart disease: treatments following myocardial infarction. lAMA 1988; 260:2088-93 17 Radzik D, Talajic M, Roy D, Hii ITY, Thibault B, Theroux P, et aI. Have changes in the management of acute myocardial infarction between 1980 and 1990 reduced late arrhythmic complications [abstract]? 1 Am Coli Cardiol 1993; 21(suppl A):43A
Thrombosis of Prosthetic Valves The George Gabor Foundation will present this conference in Budapest, Hungary, September 15-17 at the Hotel Thermal Helia. For information, contact Natyas Jektaum MD, Hungarian Institute of Cardiology, PO Box 88, H-1450 Budapest, Hungary.
1892
Reviewof 2,070 Patients WIth Acuts MyocardIaIlnIarctIon (TsuyuId at B1)