Interheart risk modifiable factors in micardio infraction 2004

Articles
Introduction
Worldwide, cardiovascular disease is estimated to be the
leading cause of death and loss of disability-adjusted life
years.1
Although age-adjusted cardiovascular death rates
have declined in several developed countries in past
decades, rates of cardiovascular disease have risen
greatly in low-income and middle-income countries,1,2
with about 80% of the burden now occurring in these
countries. Effective prevention needs a global strategy
based on knowledge of the importance of risk factors for
cardiovascular disease in different geographic regions
and among various ethnic groups.
Current knowledge about prevention of coronary heart
disease and cardiovascular disease is mainly derived from
studies done in populations of European origin.2
Researchers are unsure to what extent these findings
apply worldwide. Some data suggest that risk factors for
coronary heart disease vary between populations—eg,
lipids are not associated with this disorder in south
Asians,3
and increases in blood pressure might be more
important in Chinese people.4
Even if the association of a
risk factor with coronary heart disease is similar across
populations, prevalence of this factor might vary,
resulting in different population attributable risks
(PAR)—eg, serum cholesterol might be lower in Chinese
populations.4
On the other hand, these apparent
variations between ethnic populations could be
attributable to differences between studies in their design
and analysis, information obtained, and small sample
sizes.
To clarify whether the effects of risk factors vary in
different countries or ethnic groups, a large study
undertaken in many countries—representing different
regions and ethnic groups and using standardised
methods—is needed, with the aim to investigate the
relation between risk factors and coronary heart
disease. Such a study could also estimate the
importance of known risk factors on the PAR for acute
myocardial infarction. This aim, however, needs either
very large cohort trials or case-control studies with
many events—eg, several thousands of cases of
myocardial infarction in whom all (or most) currently
Lancet 2004; 364: 937–52
Published online
September 3, 2004
http://image.thelancet.com/
extras/04art8001web.pdf
See Comment page 912
*Listed at end of report.
Population Health Research
Institute, Hamilton General
Hospital, 237 Barton Street
East, Hamilton, Ontario,
Canada L8L 2X2
(Prof S Yusuf DPhil,
S Ôunpuu PhD, S Hawken MSc,
T Dans MD, A Avezum MD,
F Lanas MD, M McQueen FRCP,
A Budaj MD, P Pais MD,
J Varigos BSc, L Lisheng MD)
Correspondence to:
Prof Salim Yusuf
yusufs@mcmaster.ca
www.thelancet.com Vol 364 September 11, 2004 937
Effect of potentially modifiable risk factors associated with
myocardial infarction in 52 countries (the INTERHEART
study): case-control study
Salim Yusuf, Steven Hawken, Stephanie Ôunpuu, Tony Dans, Alvaro Avezum, Fernando Lanas, Matthew McQueen, Andrzej Budaj, Prem Pais,
John Varigos, Liu Lisheng, on behalf of the INTERHEART Study Investigators*
Summary
Background Although more than 80% of the global burden of cardiovascular disease occurs in low-income and
middle-income countries, knowledge of the importance of risk factors is largely derived from developed countries.
Therefore, the effect of such factors on risk of coronary heart disease in most regions of the world is unknown.
Methods We established a standardised case-control study of acute myocardial infarction in 52 countries,
representing every inhabited continent. 15152 cases and 14820 controls were enrolled. The relation of smoking,
history of hypertension or diabetes, waist/hip ratio, dietary patterns, physical activity, consumption of alcohol, blood
apolipoproteins (Apo), and psychosocial factors to myocardial infarction are reported here. Odds ratios and their
99% CIs for the association of risk factors to myocardial infarction and their population attributable risks (PAR) were
calculated.
Findings Smoking (odds ratio 2·87 for current vs never, PAR 35·7% for current and former vs never), raised
ApoB/ApoA1 ratio (3·25 for top vs lowest quintile, PAR 49·2% for top four quintiles vs lowest quintile), history of
hypertension (1·91, PAR 17·9%), diabetes (2·37, PAR 9·9%), abdominal obesity (1·12 for top vs lowest tertile and
1·62 for middle vs lowest tertile, PAR 20·1% for top two tertiles vs lowest tertile), psychosocial factors (2·67, PAR
32·5%), daily consumption of fruits and vegetables (0·70, PAR 13·7% for lack of daily consumption), regular alcohol
consumption (0·91, PAR 6·7%), and regular physical activity (0·86, PAR 12·2%), were all significantly related to
acute myocardial infarction (p<0·0001 for all risk factors and p=0·03 for alcohol). These associations were noted in
men and women, old and young, and in all regions of the world. Collectively, these nine risk factors accounted for
90% of the PAR in men and 94% in women.
Interpretation Abnormal lipids, smoking, hypertension, diabetes, abdominal obesity, psychosocial factors, consumption
of fruits, vegetables, and alcohol, and regular physical activity account for most of the risk of myocardial infarction
worldwide in both sexes and at all ages in all regions. This finding suggests that approaches to prevention can be based
on similar principles worldwide and have the potential to prevent most premature cases of myocardial infarction.

Articles
known risk factors are measured. We judged the latter
most practical.
INTERHEART is a large, international, standardised,
case-control study, designed as an initial step to assess
the importance of risk factors for coronary heart disease
worldwide (slides available at http://www.phri.ca/
interheart).5
We aimed to include about 15 000 cases
and a similar number of controls from 52 countries,
representing all inhabited continents. Specific
objectives are to determine the strength of association
between various risk factors and acute myocardial
infarction in the overall study population and to
ascertain if this association varies by geographic region,
ethnic origin, sex, or age. A key secondary objective is to
estimate the PAR for risk factors and their
combinations in the overall population and in various
subgroups. This report focuses on the association of
nine easily measured protective or risk factors
(smoking, lipids, self-reported hypertension or
diabetes, obesity, diet, physical activity, alcohol
consumption, and psychosocial factors) to first
myocardial infarction.
Methods
Participants
Study participants were recruited from 262 centres from
52 countries in Asia, Europe, the Middle East, Africa,
Australia, North America, and South America (web-
table 1; http://image.thelancet.com/extras/04art8001web
table1.pdf). The national coordinator selected centres
within every country on the basis of feasibility. To identify
first cases of acute myocardial infarction, all patients
(irrespective of age) admitted to the coronary care unit or
equivalent cardiology ward, presenting within 24 h of
symptom onset, were screened. Cases were eligible if they
had characteristic symptoms plus electrocardiogram
changes indicative of a new myocardial infarction
(webappendix 1; http://image.thelancet.com/extras/
04art8001webappendix1.pdf).
At least one age-matched (up to 5 years older or
younger) and sex-matched control was recruited per
case, using specific criteria. Exclusion criteria for
controls were identical to those described for cases, with
the additional criterion that controls had no previous
diagnosis of heart disease or history of exertional chest
pain. The overall median interval from recruitment of
cases to inclusion of controls was 1·5 months. Hospital-
based controls (58%) were individuals who had a wide
range of disorders unrelated to known or potential risk
factors for acute myocardial infarction and were
admitted to the same hospital as the matching case.
Community-based controls (36%) were attendants or
relatives of a patient from a non-cardiac ward or an
unrelated (not first-degree relative) attendant of a cardiac
patient. In the remaining controls, 3% were from an
undocumented source and 3% were recruited through
the WHO MONICA study.6
Procedures
Structured questionnaires were administered and
physical examinations were undertaken in the same
manner in cases and controls. Information about
demographic factors, socioeconomic status (education,
income), lifestyle (smoking, leisure time, physical activity,
and dietary patterns), personal and family history of
cardiovascular disease, and risk factors (hypertension,
diabetes mellitus) was obtained. Psychosocial factors
(depression, locus of control, perceived stress, and life
events) were systematically recorded and integrated into
one score: details are provided in the accompanying
paper.7
Height, weight, waist and hip circumferences, and
heart rate were determined by a standardised protocol.
Waist and hip circumferences were measured with a non-
stretchable standard tape measure: waist measurements
were obtained over the unclothed abdomen at the
narrowest point between the costal margin and iliac crest,
and hip circumferences over light clothing at the level of
the widest diameter around the buttocks. Although blood
pressure at the time of examination was recorded in both
cases and controls, the levels in cases would be
systematically affected by the myocardial infarction and
treatments—eg, ␤ blockers, nitrates, and angiotensin-
converting-enzyme inhibitors—that could lower blood
pressure. Therefore, only self-reported history of
hypertension is used in the analysis.
Non-fasting blood samples (20 mL) were drawn from
every individual and centrifuged within 2 h of
admission, separated into six equal volumes, and
frozen immediately at –20°C or –70°C after processing.
Centres were instructed to draw blood from cases
within 24 h of symptom onset. However, because of
delays in patient presentation, especially in some low-
income countries, blood samples could only be
obtained within 24 h in two-thirds of cases. Samples
were shipped in nitrogen vapour tanks by courier from
every site to a blood storage site, where they were stored
at –160°C in liquid nitrogen (Hamilton, Canada) or at
–70°C (India and China). Blood samples from all
countries other than China were analysed in Hamilton
for total cholesterol, HDL cholesterol, and
apolipoproteins B (ApoB) and A1 (ApoA1).
Immunoturbidimetric assays were used to measure
apolipoprotein concentrations (Roche/Hitachi 917
analyser with Tina-quant ApoB version 2 and ApoA1
version 2 kits; Roche Diagnostics, Mannheim, Germany).
The ApoB method was standardised against the IFCC
SP3–07 reference standard8
and the ApoA1 method
against the IFCC SP1–01 reference preparation.9
The
same measurement kits and a Roche/Hitachi 911
analyser were used in Beijing, China. Both laboratories
measured the same lot numbers of Precinorm and
Precipath controls (Roche Diagnostics) in every run, and
in every patient sample analysis run in China, two study
patients and two serum reference pool samples (pool A
and B) were measured that had previously been analysed
See Articles page 953
938 www.thelancet.com Vol 364 September 11, 2004

Articles
in the central core laboratory in Canada. Because
apolipoprotein concentrations are not affected by the
fasting status of the individual (unlike calculated LDL),
we used the ApoB/ApoA1 ratio as an index of abnormal
lipids in the current analysis.10
Moreover, this ratio was
predictive of myocardial infarction in subsets of patients
(<12 h, 12–24 h, and >24 h after symptoms) in the
present study (data not shown). Detailed information on
lipoprotein fractions will be reported separately.
All data were transferred to the Population Health
Research Institute, McMaster University, and Hamilton
Health Sciences, Canada, where quality-control checks
and statistical analyses were done. Data on smoking
were missing in 1·1% of participants, hypertension in
0·6%, diabetes in 0·7%, psychosocial variables in 11%,
physical activity in 1·1%, diet in 2·1%, and waist and hip
measurements in 3·5%. Blood samples were available in
21508 (79%) of 27098 cases and controls.
INTERHEART was approved by appropriate
regulatory and ethics committees in all participating
countries and centres. All participants provided
informed consent before taking part in the study.
We defined current smokers as individuals who smoked
any tobacco in the previous 12 months and included those
who had quit within the past year. Former smokers were
defined as those who had quit more than a year earlier.
For waist/hip ratio, tertiles were calculated separately for
men and women based on the overall control data. The
cutoffs used were 0·90 and 0·95 in men and 0·83 and
0·90 in women, to divide participants into thirds. Cutoffs
for ApoB/ApoA1 ratios (deciles and quintiles) were
derived from all controls (men and women). Region-
specific cutoffs did not alter the results. Individuals were
judged to be physically active if they were regularly
involved in moderate (walking, cycling, or gardening) or
strenuous exercise (jogging, football, and vigorous
swimming) for 4 h or more a week. Regular alcohol use
was defined as consumption three or more times a week.
The combined psychosocial index was devised with a
combination of the parameter estimates from the
completely adjusted multivariate logistic regression
model. The score was based on a combination of
depression versus none, stress at work or at home (general
stress variable) versus none, moderate or severe financial
stress versus minimal or none, one or more life events
versus none, and a locus of control score in the lower three
quartiles versus the top quartile of the distribution.
Statistical analysis
Simple associations were assessed with frequency tables
and Pearson’s ␹2
tests for two independent proportions.
For comparison of prevalence across distinct
subgroups—eg, by region, country, or ethnic group—
potential differences in age structure of the populations
were accounted for by direct standardisation of the
frequencies to the overall INTERHEART age
distribution with a five-level age-stratification factor
(<45, 46–55, 56–65, 66–70, >70).11
Means and medians
were calculated to summarise continuous effects and
were compared by t tests or appropriate non-parametric
tests when distributional assumptions were in doubt.
When data have been categorised by tertiles, quintiles, or
deciles, these were based on the overall control data. For
waist/hip ratio, sex-specific cutoffs were used. For
protective factors (exercise, diet, and alcohol), the PAR is
calculated for the group without the exposure.
The findings presented are for models fitted with
unconditional logistic regression, adjusted for the
matching criteria, for two reasons. First, unmatched
analyses were used because for 14% (1763/12461) of cases
of myocardial infarction and 5% (738/14637) of controls,
perfect matching was not possible. Undertaking a strict
matched analysis would mean relevant loss of information
because of the exclusion of these participants. Moreover,
when data on a risk factor were missing in a case or
control, the entire pair would be excluded from all
analyses. Therefore, we widened the age-matching criteria
and used frequency matching of cases and controls, using
age and sex strata. Second, there was general agreement
for key results among the many methods compared
(conditional logistic regression, mixed models, and
unconditional logistic regression, with adjustment for
matching criteria). Estimated odds ratios and CIs
calculated with the different methods were within 5% of
each other, with a slight attenuation of effect estimates in
the unconditional versus conditional models (webtable 2;
http://image.thelancet.com/extras/04art8001webtable2.
pdf).11
Hence, findings presented are adjusted for age, sex,
geographic region, and potential confounders, and should
be interpreted as providing a slight underestimation of
effect sizes for most comparisons.
Adjusted odds ratios for combinations of risk factors can
be derived from their respective model coefficients in the
multivariate logistic model. By summation of model
coefficients and taking the antilog (panel) the combined
effect of combinations of exposures can be estimated.
Estimates of odds ratios and accompanying 99% CIs are
presented for every risk factor and their combinations.
Statistical analyses and graphics were produced
with SAS version 8.2 (SAS, Cary, NC, USA) and
S-Plus version 6 (Insightful, Seattle, WA, USA). All
statistical tests of hypotheses are two-sided. PARs and
99% CIs were calculated for various risk factors in the
study by a method based on unconditional logistic
regression.12
The PARs presented are adjusted for
confounders in a similar manner to the corresponding
logistic regression models for odds ratio estimates and,
where indicated, are stratified by subgroups of interest.
PAR estimates were calculated by the interactive risk
attributable program software (US National Cancer
Institute, 2002).13
For a simple dichotomous exposure and disease, and
no adjustment for confounding, the usual formula for
PAR was used (panel).12
PAR adjusted for confounding

Articles
is also shown in the panel. For variance estimates, the
reader is referred to Benichou and Gail15
since the
derivations and formulae are complex. CI calculations
were based on this method using a logit transformation
approach, apart from when PAR estimates were
negative, in which case conventional Wald type CIs
were used.
Role of the funding source
The sponsors of the study had no role in study design,
data collection, data analysis, data interpretation, or
writing of the report. The corresponding author had full
access to all the data in the study and had final
responsibility for the decision to submit for publication.
Results
Between February, 1999, and March, 2003, 15152 cases
and 14820 controls were enrolled. 1531 cases were
diagnosed as having unstable angina, 260 had insufficient
data, 205 did not have coronary artery disease, and 695
had a previous myocardial infarction. For 74 controls data
were missing and 109 had previous coronary heart
disease. Therefore, 12461 cases and 14637 controls are
included in the analysis. Table 1 shows the distribution of
participants by region and ethnic origin. 9459 cases (76%)
and 10851 controls (74%) were male.
Table 2 shows the median age of presentation of cases.
The overall median age of cases with first acute
myocardial infarction is about 9 years lower in men than
in women in all regions of the world. However, the
proportion of male cases was highest in countries with a
younger age of presentation of acute myocardial
infarction—eg, 85% of cases in south Asia and 86% in the
Middle East were male compared with 74% in western
Europe, 68% in central and eastern Europe, and 70% in
China. Among regions, striking variations were noted in
the age of first presentation of acute myocardial
infarction, with the youngest patients in south Asia
(median age 53 years) and the Middle East (51 years), and
the oldest patients in western Europe, China, and Hong
Kong (63 years). The highest proportion of cases with first
acute myocardial infarction at age 40 years or younger was
in men from the Middle East (12·6%), Africa (10·9%),
and south Asia (9·7%) and the lowest proportion was in
women from China and Hong Kong (1·2%), South
America (1·0%), and central and eastern Europe (0·9%).
Overall effect of risk factors
Table 3 provides the overall odds ratios for individual risk
factors adjusted for age, sex, smoking status, and region
and by multivariate adjustment for all risk factors. All
risk factors were significantly (p<0·0001) related to acute
myocardial infarction, except alcohol, which had a
weaker association (p=0·03). After multivariate analysis,
current smoking and raised ApoB/ApoA1 ratio (top vs
lowest quintile) were the two strongest risk factors,
followed by history of diabetes, hypertension, and
psychosocial factors (table 3). Body-mass index was
related to risk of myocardial infarction, but this relation
was weaker than that of abdominal obesity (waist/hip
ratio), with body-mass index becoming non-significant
with the inclusion of waist/hip ratio in the multivariate
model (data not shown). Before multivariate adjustment,
abdominal obesity (top vs lowest tertile) doubled the risk
of acute myocardial infarction, but the effects were
substantially diminished after adjustment for other risk
factors, especially apolipoproteins. Daily consumption of
fruits or vegetables, moderate or strenuous physical
exercise, and consumption of alcohol three or more
times per week, were protective (table 3).
Panel: Formulae
X is the exposure level, C is the confounder level, and D is disease status (D=0, disease is
absent; D=1, disease is present).
Cases (n=12 461) Controls (n=14 637)
Geographic region
Western Europe 664 767
Central and eastern Europe 1727 1927
Middle East 1639 1786
Africa 578 789
South Asia 1732 2204
China and Hong Kong 3030 3056
Southeast Asia and Japan 969 1199
Australia and New Zealand 589 681
South America and Mexico 1237 1888
North America 296 340
Ethnic origin
European 3314 3710
Chinese 3130 3167
South Asian 2171 2573
Other Asian 871 1073
Arab 1306 1479
Latin American 1141 1834
Black African 157 369
Coloured African 311 339
Other 60 93
Table 1: Distribution of study population
where Pr(E) is probability of exposure to the risk factor and R is the relative risk of the
disease in exposed versus unexposed individuals.
where
and

Articles
A strong and graded relation was noted between
numbers smoked and risk of myocardial infarction, with
the risk increasing at every increment, so that individuals
smoking greater than 40 cigarettes per day had an odds
ratio of 9·16 (99% CI 6·18–13·58; figure 1). The
ApoB/ApoA1 ratio also showed a graded relation with
myocardial infarction risk, with no evidence of a threshold,
with an odds ratio of 4·73 (99% CI 3·93–5·69) for the top
versus the lowest decile of ApoB/ApoA1 ratio (figure 1).
Cumulative effect of risk factors
Figure 2 shows the effect of multiple risk factors on
increased risk of myocardial infarction. Together, current
smoking, hypertension, and diabetes increased the odds
ratio for acute myocardial infarction to 13·01 (99% CI
10·69–15·83) compared to those without these risk
factors, and they accounted for 53% of the PAR of acute
myocardial infarction. Addition of ApoB/ApoA1 ratio (top
vs lowest quintile) increased the odds ratio to
Overall Men Women
Number Median age (IQR) % <40 years (n) Number Median age (IQR) % <40 years (n) Number Median age (IQR) % <40 years (n)
Geographic region
Western Europe 664 63 (54–72) 2·7 (18) 493 61 (53–70) 2·8 (14) 171 68 (59–76) 2·3 (4)
Central and eastern Europe 1727 62 (52–70) 2·9 (51) 1173 59 (50–68) 3·9 (46) 554 68 (59–74) 0·9 (5)
North America 296 59 (50–71) 4·0 (12) 210 58 (49–68) 3·3 (7) 86 64 (52–75) 5·8 (5)
South America and Mexico 1237 60 (51–70) 3·4 (42) 926 59 (50–68) 4·2 (39) 311 65 (56–73) 1·0 (3)
Australia and New Zealand 589 60 (51–69) 5·3 (31) 464 58 (50–67) 5·6 (26) 125 66 (59–74) 4·0 (5)
Middle East 1639 51 (45–59) 11·2 (184) 1410 50 (44–57) 12·6 (177) 229 57 (50–65) 3·1 (7)
Africa 578 54 (47–62) 9·7 (56) 385 52 (46–61) 10·9 (42) 193 56 (49–65) 7·3 (14)
South Asia 1732 53 (46–61) 8·9 (54) 1480 52 (45–60) 9·7 (143) 252 60 (50–66) 4·4 (11)
China and Hong Kong 3030 63 (53–70) 4·5 (135) 2131 60 (50–68) 5·8 (124) 899 67 (62–72) 1·2 (11)
Southeast Asia and Japan 969 57 (49–65) 7·0 (68) 787 55 (47–64) 8·3 (65) 182 63 (56–68) 1·7 (3)
Ethnic origin
European 3314 62 (52–71) 3·2 (107) 2371 59 (51–69) 3·8 (89) 943 68 (58–75) 1·9 (18)
Chinese 3130 63 (53–70) 4·4 (139) 2217 60 (50–68) 5·8 (128) 913 67 (61–72) 1·2 (11)
South Asian 2171 52 (45–60) 10·6 (231) 1889 50 (45–60) 11·7 (220) 282 60 (51–66) 3·9 (11)
Other Asian 871 57 (48–65) 7·0 (61) 705 55 (47–64) 8·2 (58) 166 63 (56–68) 1·8 (3)
Arab 1306 53 (46–60) 9·0 (118) 1083 52 (45–59) 10·3 (111) 223 57 (50–65) 3·1 (7)
Latin American 1141 60 (51–69) 3·7 (42) 854 58 (50–67) 4·5 (38) 287 64 (55–72) 1·4 (4)
Black African 157 52 (46–61) 14·0 (22) 98 52 (46–59) 17·4 (17) 59 54 (48–67) 8·5 (5)
Coloured African 311 54 (47–63) 8·7 (27) 196 52 (46–62) 9·7 (19) 115 58 (49–65) 7·0 (8)
Other 60 57 (48–64) 6·7 (4) 46 53 (48–62) 6·5 (3) 14 63 (59–73) 7·1 (1)
Overall 12 461 58 (49–67) 6·0 (751) 9459 56 (48–65) 7·2 (683) 3002 65 (56–72) 2·3 (68)
Table 2: Median age (years) of presentation of cases
Prevalence Odds ratio (99% CI) adjusted for age, PAR (99% CI) Odds ratio (99% CI) adjusted additionally PAR 2 (99% CI)
sex, and smoking (OR 1) for all other risk factors (OR 2)
Controls (%) Cases (%)
Risk factor
Current smoking* 26·76 45·17 2·95 (2·72–3·20) – 2·87 (2·58–3·19) –
Current and former smoking* 48·12 65·19 2·27 (2·11–2·44) 36·4% (33·9–39·0) 2·04 (1·86–2·25) 35·7% (32·5–39·1)
Diabetes 7·52 18·45 3·08 (2·77–3·42) 12·3% (11·2–13·5) 2·37 (2·07–2·71) 9·9% (8-5–11·5)
Hypertension 21·91 39·02 2·48 (2·30–2·68) 23·4% (21·7–25·1) 1·91 (1·74–2·10) 17·9% (15·7–20·4)
Abdominal obesity (2 vs 1)† 33·40 30·21 1·36 (1·24–1·48) – 1·12 (1·01–1·25) –
Abdominal obesity (3 vs 1)† 33·32 46·31 2·24 (2·06–2·45) 33·7% (30·2–37·4) 1·62 (1·45–1·80) 20·1% (15·3–26·0)
All psychosocial‡ – – 2·51 (2·15–2·93) 28·8% (22·6–35·8) 2·67 (2·21–3·22) 32·5% (25·1–40·8)
Vegetables and fruit daily* 42·36 35·79 0·70 (0·64–0·77) 12·9% (10·0–16·6) 0·70 (0·62–0·79) 13·7% (9·9–18·6)
Exercise* 19·28 14·27 0·72 (0·65–0·79) 25·5% (20·1–31·8) 0·86 (0·76–0·97) 12·2% (5·5–25·1)
Alcohol intake* 24·45 24·01 0·79 (0·73–0·86) 13·9% (9·3–20·2) 0·91 (o·82–1·02) 6·7% (2·0–20·2)
ApoB/ApoA1 ratio (2 vs 1)§ 19·99 14·26 1·47 (1·28–1·68) – 1·42 (1·22–1·65) –
ApoB/ApoA1 ratio (5 vs 1)§ 20·00 33-49 3·87 (3·39–4·42) 54·1% (49·6–58·6) 3·25 (2·81–3·76) 49·2% (43·8–54·5)
All above risk factors combined¶ – – 129·20 (90·24–184·99) 90·4% (88·1–92·4) 129·20 (90·24–184·99) 90·4% (88·1–92·4)
The median waist /hip ratio was 0·93 in cases and 0·91 in controls (p<0·0001), and the median ApoB/ApoA1 ratio was 0·85 in cases and 0·80 in controls (p<0·0001). Percentage of controls with four or five factors positive is
22·2% compared with 29·2% in cases. *PARs for smoking, abdominal obesity, and ApoB/ApoA1 ratio are based on a comparison of all smokers vs never, top two tertiles vs lowest tertile, and top four quintiles vs lowest quintile.
For protective factors (diet, exercise, and alcohol), PARs are provided for the group without these factors. †Top two tertiles vs lowest tertile. ‡A model-dependent index combining positive exposure to depression, perceived
stress at home or work (general stress), low locus of control, and major life events, all referenced against non-exposure for all five factors. §Second, third, fourth, or fifth quintiles vs lowest quintile. ¶The model is saturated, so
adjusted and unadjusted estimates are identical for all risk factors. The odds ratio of 129·20 is derived from combining all risk factors together, including current and former smoking vs never smoking, top two tertiles vs lowest
tertile of abdominal obesity, and top four quintiles vs lowest quintile of ApoB/ApoA1. If, however, the model includes only current smoking vs never smoking, the top vs lowest tertile for abdominal obesity, and the top vs lowest
quintile for ApoB/ApoA1, the odds ratio for the combined risk factors increases to 333·7 (99% CI 230·2–483·9).
Table 3: Risk of acute myocardial infarction associated with risk factors in the overall population

Articles
42·3 (33·2–54·0), and the PAR for these four risk factors
together (top four quintiles of ApoB/ApoA1 ratio vs lowest
quintile) was 75·8% (99% CI 72·7–78·6). Addition of
abdominal obesity (top two tertiles vs lowest tertile)
further increased the PAR to 80·2% (77·5–82·7).
Figure 3 shows the effects of multiple risk factors on
reduced risk of acute myocardial infarction associated
with healthy lifestyles. Daily consumption of fruit and
vegetables and regular physical activity conferred an
odds ratio of 0·60 (99% CI 0·51–0·71). Further, if an
individual avoided smoking, the odds ratio would be
0·21 (0·17–0·25; figure 3), suggesting that
modification of these aspects of lifestyle could
potentially reduce the risk of an acute myocardial
infarction by more than three-quarters compared with a
smoker with a poor lifestyle.
Incorporation of all nine independent risk factors
(current or former smoking, history of diabetes or
hypertension, abdominal obesity, combined psycho-
social stressors, irregular consumption of fruits and
vegetables, no alcohol intake, avoidance of any regular
exercise, and raised plasma lipids) indicates an odds
ratio of 129·20 (99% CI 90·24–184·99; table 3),
compared with not having any of these risk factors.
Substituting the odds ratios for current smoking, the
extremes of abdominal obesity (top vs lowest tertile) and
ApoB/ApoA1 ratio (top vs lowest quintile) increases the
combined effect of all nine risk factors to 333·7 (99% CI
230·2–483·9; figure 2). This represents a PAR of 90·4%
(99% CI 88·1–92·4), suggesting that these risk factors
account for most of the risk of acute myocardial
infarction in our study population. In view of the overlap
in the effect of the nine risk factors, most of the PAR
could be accounted for by a combination of various risk
factors, as long as they included smoking and the
ApoB/ApoA1 ratio (PAR for their combination is 66·8%
[99% CI 62·8–70·6]). The estimate of the combined
effect of all nine risk factors is derived from a model,
since very few individuals had zero risk factors or all
nine risk factors. However, confidence that the majority
of risk is indeed accounted for by these risk factors is
lent support by the fact that of the 18708 individuals
with complete data on all risk factors, 43 controls and
24 cases had no risk factors and 49 cases and 11 controls
had eight or more. Also, just five risk factors (smoking,
lipids, hypertension, diabetes, and obesity), which a
large proportion of individuals had, accounted for about
80% of the PAR.
Risk in men and women
Figure 4 presents odds ratios and PARs for risk of acute
myocardial infarction in men and women. Similar odds
ratios were recorded in women and men for the
association of acute myocardial infarction with smoking,
raised lipids, abdominal obesity, composite of
psychosocial variables, and vegetable and fruit
consumption. However, the increased risk associated
with hypertension and diabetes, and the protective effect
of exercise and alcohol, seemed to be greater in women
then in men (figure 4).
Table 4 also shows PARs by sex for the various risk
factors, adjusted for age and region only and the fully
adjusted model. In men, smoking was associated with
42·7% of the PAR for acute myocardial infarction
compared with 14·8% in women in the fully adjusted
model. Abnormal lipids had the highest PAR in both
men (49·5%) and women (47·1%), with high
contributions from psychosocial risk factors (28·8% vs
45·2%) and abdominal obesity (19·7% vs 18·7%).
Hypertension contributed to PAR in women to a greater
extent (29·0%) compared with men (14·9%), partly
because of a higher prevalence of hypertension in
women who were about a decade older. Collectively, all
nine risk factors accounted for 90% of the PAR in men
and 94% in women (table 4).
Numbers smoked per day
Number of controls
Number of cases
Odds ratio
Oddsratio(99%CI)
Never 1–5 6–10 11–15 16–20 21–25 26–30 31–40 у41
7489 727 1031 446 1058 96 230 168 56
4223 469 1021 623 1832 254 538 459 218
1 1·38 2·10 2·99 3·83 5·80 5·26 6·34 9·16
16
8
4
2
1
ApoB/A1 ratio (deciles)
Number of controls
Number of cases
Median
Oddsratio(99%CI)
1 2 3 4 5 6 7 8 10
1210 1206 1208 1207 1210 1209 1207 1208 1209
435 496 610 720 790 893 1063 1196 1757
0·43 0·53 0·60 0·66 0·72 0·78 0·85 0·93 1·28
1208
1366
1·04
8
4
2
1
0·75
9
Figure 1: Odds of myocardial infarction according to number of cigarettes smoked and ApoB/ApoA1 ratio
Note the doubling scale on the y axis for both figures.

Articles
Risk by age
Smoking, adverse lipid profile, hypertension, and
diabetes had a greater relative effect on risk of acute
myocardial infarction in younger than older individuals
(table 5). Overall, abnormal lipids was the most
important risk factor with respect to PAR in both young
and old individuals (table 5). Collectively, the nine risk
factors accounted for a significantly greater (p<0·0001)
PAR in younger than older individuals; these patterns
were consistent in males and females.
Regional and ethnic variations in importance of risk
factors
When the odds ratio (adjusted for age, sex, smoking,
and geographic region) for association of acute
myocardial infarction with a risk factor is around 2 or
more, eg, for smoking, lipids, hypertension, diabetes,
abdominal obesity, and the combined psychosocial
index, subgroup analyses are likely to be fairly robust.
We recorded a clear, significant, and consistent excess
risk of acute myocardial infarction associated with
these risk factors in most regions of the world and in
every ethnic group (figures 5–10). By contrast, when
odds ratios were weaker (0·70–1·50; alcohol
consumption, exercise, or diet), greater variability was
noted across regions (data not shown). This apparent
variability could be attributable to chance, because
subgroup analyses are likely to be less reliable when
smaller overall differences are subdivided across
multiple subsets of the populations. Similar results
were noted for analyses across various subgroups
defined by ethnic origin, with consistent and clear
excess risks being reported for tobacco use, abnormal
lipids, history of hypertension, diabetes, and
abdominal obesity (data not shown).
Population attributable risk by geographic region
Table 4 also presents overall PARs and values by sex
across different geographic regions. In all regions, the
nine risk factors account for between three-quarters
and virtually all the PAR for acute myocardial
infarction. The relative importance of every risk factor
varied, and was largely related to its prevalence.
However, raised lipids, smoking, and psychosocial
factors were the most important risk factors in all
regions in the world. It is noteworthy that in western
Europe, North America, and Australia and New
Zealand (representing high-income countries) and
southeast Asia (mostly middle-income countries),
abdominal obesity was associated with a PAR greater
than that associated with smoking. A similar pattern
was seen for Africa, but most of our data are drawn
from South Africa, which is a middle-income country.
However, obesity was less important in other parts of
the world, where it is less prevalent. For example,
obesity accounted for only 5·5% of the PAR in China
compared with 35·8% for smoking (where 41% of
male and 4% of female controls smoked). Subdividing
the population by ethnic origin, these nine risk factors
accounted for a very high proportion of the PAR in
every ethnic group (Europeans, 86%; Chinese, 90%;
south Asians, 92%; black Africans, 92%; Arabs, 93%;
and Latin Americans, 90%).
Figure 2: Risk of acute myocardial infarction associated with exposure to multiple risk factors
Smk=smoking. DM=diabetes mellitus. HTN=hypertension. Obes=abdominal obesity. PS=psychosocial. RF=risk
factors. Note the doubling scale on the y axis. The odds ratios are based on current vs never smoking, top vs lowest
tertile for abdominal obesity, and top vs lowest quintile for ApoB/ApoA1. If these three are substituted by current
and former smoking, top two tertiles for abdominal obesity and top four quintiles for ApoB/ApoA1, then the odds
ratio for the combined risk factor is 129·20 (99% CI 90·24–184·99).
Figure 3: Reduced risk of acute myocardial infarction associated with various risk factors
Smk=smoking. Fr/vg=fruits and vegetables. Exer=exercise. Alc=alcohol. Note the doubling scale on the y axis. Odds
ratios are adjusted for all risk factors.
2·9
(2·6–3·2)
2·4
(2·1–2·7)
1·9
(1·7–2·1)
3·3
(2·8–3·8)
13·0
(10·7–15·8)
42·3
(33·2–54·0)
68·5
(53·0–88·6)
182·9
(132·6–252·2)
333·7
(230·2–483·9)
Oddsratio(99%CI)
Smk
(1)
DM
(2)
HTN
(3)
ApoB/A1
(4)
1+2+3 All 4 All RFs
512
32
64
128
256
16
4
8
2
1
+Obes +PS
0·35
(0·31–0·39)
0·70
(0·62–0·79)
0·86
(0·76–0·97)
0·91
(0·82–1·02)
0·24
(0·21–0·29)
0·21
(0·17–0·25)
0·19
(0·15–0·24)
Oddsratio(99%CI)
No
Smk
Fr/vg Alc
Risk factor (adjusted for all others)
No Smk
+Fr/Vg
+Ex +Alc
1·0
0·5
0·25
0·125
0·0625
Exer

Articles
Consistency of results
Subgroup analyses with both types of controls (hospital-
based and community-based) showed consistent odds
ratios for current smoking (hospital-based 3·1 vs
community-based 2·8), for the top quintile versus lowest
quintile of lipids (4·2 vs 3·9), for diabetes (2·7 vs 3·4),
for hypertension (2·1 vs 3·0), for abdominal obesity (1·7
vs 1·9), for psychosocial factors (1·6 vs 1·5), for
consumption of fruits (0·78 vs 0·93) and vegetables
(0·78 vs 0·83), for regular physical activity (0·79 vs 0·79),
and for alcohol use (0·79 vs 0·86).
583 cases of acute myocardial infarction subsequently
died in hospital. Odds ratios for fatal myocardial
infarction associated with various risk factors were
similar to those overall—smoking (2·1 for fatal
myocardial infarction vs 3·0 overall), diabetes (4·0 vs
3·1), hypertension (2·4 vs 2·5), abdominal obesity (1·5
vs 2·2), and lipids (2·6 vs 3·9).
Family history
Family history of coronary heart disease was associated
with an odds ratio of 1·55 (99% CI 1·44–1·67),
adjusted for age, sex, smoking, and geographic region.
Adjustments for the nine previously described risk
factors slightly reduced the odds ratio to 1·45
(1·31–1·60). The PAR was 12·0% (99% CI
9·2%–15·1%), which fell to 9·8% (7·6–12·5) after full
adjustment. However, when family history is added to
the information from other nine risk factors, the
overall PAR rose from 90·4% to only 91·4%,
indicating that although family history is an
independent risk factor for myocardial infarction, most
of the associated risk burden can be accounted for
through the other risk factors studied. Family history
seemed to be slightly more important in young (PAR
14·8% [11·7–18·5]) compared with old individuals
(10·4% [8·3–13·0]).
Repeat measures
Repeat measures of risk factors were made in 279 controls
at a median interval of 409 days. The agreement rates for
smoking (Cohen’s kappa16
␬=0·94), history of diabetes
(␬=0·90), ApoB/ApoA1 (intraclass correlation=0·74),
hypertension (␬=0·82), depression (␬=0·44), abdominal
obesity (intraclass correlation=0·68), regular physical
activity (␬=0·56), and consumption of fruits (␬=0·66),
vegetables (␬=0·52), and alcohol (␬=0·52) were high to
moderate. These data suggest that the association of
myocardial infarction with smoking and diabetes is closer
to the real effect, whereas the association of other risk
factors measured with greater variability are probably
underestimates due to regression-dilution bias.17
Risk factor Sex Control (%) Case (%) Odds ratio (99% CI) PAR (99% CI)
Current smoking F 9·3 20·1 2·86 (2·36–3·48) 15·8% (12·9–19·3)
M 33·0 53·1 3·05 (2·78–3·33) 44·0% (40·9–47·2)
Diabetes F 7·9 25·5 4·26 (3·51–5·18) 19·1% (16·8–21·7)
M 7·4 16·2 2·67 (2·36–3·02) 10·1% ( 8·9–11·4)
Hypertension F 28·3 53·0 2·95 (2·57–3·39) 35·8% (32·1–39·6)
M 19·7 34·6 2·32 (2·12–2·53) 19·5% (17·7–21·5)
Abdominal
obesity
F 33·3 45·6 2·26 (1·90–2·68) 35·9% (28·9–43·6)
M 33·3 46·5 2·24 (2·03–2·47) 32·1% (28·0–36·5)
Psychosocial
index
F 3·49 (2·41–5·04) 40·0% (28·6–52·6)
Fruits/veg
M 2·58 (2·11–3·14) 25·3% (18·2–34·0)
F 50·3 39·4 0·58 (0·48–0·71) 17·8% (12·9–24·1)
M 39·6 34·7 0·74 (0·66–0·83) 10·3% ( 6·9–15·2)
Exercise F 16·5 9·3 0·48 (0·39–0·59) 37·3% (26·1–50·0)
M 20·3 15·8 0·77 (0·69–0·85) 22·9% (16·9–30·2)
Alcohol F 11·2 6·3 0·41 (0·32–0·53) 46·9% (34·3–60·0)
M 29·1 29·6 0·88 (0·81–0·96) 10·5% (6·1–17·5)
ApoB/ApoA1 F 14·1 27·0 4·42 (3·43–5·70) 52·1% (44·0–60·2)
ratio
M 21·9 35·5 3·76 (3·23–4·38) 53·8% (48·3–59·2)
0·5 1 2 4 8 160·25
Odds ratio (99% CI)
– –
– –
Figure 4: Association of risk factors with acute myocardial infarction in men and women after adjustment for age, sex, and geographic region
For this and subsequent ﬁgures, the odds ratios are plotted on a doubling scale. Prevalence cannot be calculated for psychosocial factors because it is derived from a
model.

Articles
Discussion
Our study shows that nine easily measured and
potentially modifiable risk factors account for an
overwhelmingly large (over 90%) proportion of the risk
of an initial acute myocardial infarction. The effect of
these risk factors is consistent in men and women,
across different geographic regions, and by ethnic
group, making the study applicable worldwide. The
effect of the risk factors is particularly striking in young
men (PAR about 93%) and women (about 96%),
indicating that most premature myocardial infarction is
preventable. Worldwide, the two most important risk
factors are smoking and abnormal lipids. Together they
account for about two-thirds of the PAR of an acute
myocardial infarction. Psychosocial factors, abdominal
obesity, diabetes, and hypertension were the next most
important risk factors in men and women, but their
relative effect varied in different regions of the world.
The usual measure of obesity (body-mass index) showed
a modest relation with acute myocardial infarction but
was not significant when abdominal obesity was
included in the analysis.
Both smoking and apolipoproteins showed a graded
relation with the odds of a myocardial infarction,
without either a threshold or a plateau in the dose
response. In particular, smoking even five cigarettes
per day increased risk. This finding suggests that there
is no safe level of smoking and that if quitting is not
possible, the risk of myocardial infarction associated
with smoking could be significantly reduced by a
reduction in the numbers smoked. The graded relation
between ApoB/ApoA1 ratio across the deciles is
consistent with findings of a Swedish study10
and
shows that most populations in the world (at least
Region Lifestyle factors Other risk factors
Smoking (%) Fruits and Exercise (%) Alcohol (%) All lifestyles (%) Hypertension (%) Diabetes (%) Abdominal All psychosocial (%) Lipids (%) All nine risk
vegetables (%) obesity (%) factors (%)
Men
Western Europe 39·0 13·3 37·7 14·1 69·6 20·5 12·8 68·6 23·7 36·7 92·0
Central and eastern Europe 40·4 7·6 –0·4 10·4 48·9 15·9 5·8 31·7 –0·9 38·7 71·9
Middle East 51·4 5·8 1·9 –2·7 50·7 5·8 13·1 23·9 37·2 72·7 94·8
Africa 45·2 –4·4 15·9 24·1 63·7 26·8 11·6 60·4 33·8 73·7 97·9
South Asia 42·0 16·0 25·5 –5·7 58·1 17·8 10·5 36·0 13·9 60·2 88·4
China 45·3 15·1 16·6 4·2 63·7 19·9 7·9 4·9 32·0 41·3 88·8
Southeast Asia and Japan 39·2 8·5 31·4 24·6 69·6 34·3 19·1 57·9 26·9 68·7 93·7
Australia and New Zealand 46·1 8·0 20·6 11·2 61·0 18·3 5·6 49·5 31·6 48·7 87·5
South America 42·4 7·1 27·6 –7·4 57·7 28·1 9·7 35·2 36·1 41·6 86·1
North America 30·9 22·4 24·7 6·6 53·9 13·9 6·1 64·7 63·7 60·0 100
Overall 1 44·0 10·3 22·9 10·5 63·8 19·5 10·1 32·1 25·3 53·8 89·8*
Overall 2 42·7 11·7 9·3 5·1 56·5 14·9 8·0 19·7 28·8 49·5 89·8*
Women
West Europe 11·1 8·4 38·3 34·2 65·2 25·9 21·0 50·6 67·1 47·9 97·1
Central and eastern Europe 13·1 12·8 42·7 29·9 65·4 42·7 15·7 20·0 15·0 26·8 86·1
Middle East 8·1 15·9 39·1 59·0 80·3 30·1 30·3 38·9 77·4 63·3 99·4
Africa 27·6 21·0 –37·9 28·8 61·2 35·1 27·5 54·6 54·9 74·6 93·3
South Asia 7·1 30·6 45·0 26·0 59·8 28·9 20·5 48·7 29·2 52·1 99·3
China 12·5 23·6 33·5 35·8 78·6 27·6 15·0 6·3 43·2 48·3 93·6
Australia and New Zealand 40·7 15·8 33·6 47·4 80·0 37·0 11·7 67·2 17·2 14·9 †
South America 25·8 5·9 27·4 44·1 71·8 47·9 22·2 63·0 37·8 59·3 96·1
North America 25·3 12·8 27·2 73·3 86·9 30·2 12·4 44·5 32·7 32·2 †
Overall 1 15·8 17·8 37·3 46·9 75·0 35·8 19·1 35·9 40·0 52·1 94·1*
Overall 2 14·8 19·1 27·1 22·1 60·6 29·0 16·1 18·7 45·2 47·1 94·1*
Men and women
West Europe 29·3 12·4 38·4 18·7 67·6 21·9 15·0 63·4 38·9 44·6 93·9
Central and eastern Europe 30·2 10·2 11·3 12·9 49·6 24·5 9·1 28·0 4·9 35·0 72·5
Middle East 45·5 7·3 4·2 –1·0 47·6 9·2 15·5 25·9 41·6 70·5 95·0
Africa 38·9 4·8 10·1 26·6 63·4 29·6 16·7 58·4 40·0 74·1 97·4
South Asia 37·4 18·3 27·1 –5·5 56·6 19·3 11·8 37·7 15·9 58·7 89·4
China 35·9 18·0 20·3 5·7 62·3 22·1 10·0 5·5 35·4 43·8 89·9
Australia and New Zealand 44·8 11·1 23·8 18·6 66·0 22·6 7·2 61·3 28·9 43·4 89·5
South America 38·3 6·6 27·6 –3·7 56·6 32·7 12·7 45·5 35·6 47·6 89·4
North America 26·1 19·8 25·6 25·5 59·9 19·0 8·0 59·5 51·4 50·5 98·7
Overall 1 36·4 12·9 25·5 13·9 62·9 23·4 12·3 33·7 28·8 54·1 90·4*
Overall 2 35·7 13·7 12·2 6·7 54·6 17·9 9·9 20·1 32·5 49·2 90·4*
PAR estimates in women in some countries are based on small numbers and so they are less reliable . Overall 1= adjusted for age, sex, and smoking, Overall 2=adjusted for all risk factors. An extended version of this table with
99% CIs is shown in webtable 3 (http://image.thelancet.com/extras/04art8001webtable3.pdf). *Saturated model, no difference between adjusted and unadjusted models. †Non-estimatable.
Table 4: PARs associated with nine risk factors in men and women by geographic region

Articles
urban) have lipid abnormalities, which increase the
risk of myocardial infarction. Since ApoB/ApoA1 ratio
was the most important risk factor in all geographic
regions in our study, a substantial modiﬁcation of its
population distribution is important for worldwide
reduction of myocardial infarction. This act will
probably need a concerted effort, including both
population-based strategies to shift the distribution
and treatments targeted at people with the greatest
abnormalities.
Both sexes Men Women
Young Old р55 years >55 years р65 years > 65 years
Odds ratios for relative effect of risk factors (99% Cl)
Lifestyle factors
Smoking 3·33 (2·86–3·87) 2·44* (2·10–2·84) 3·33 (2·80–3·95) 2·52 (2·15–2·96) 4·49 (3·11–6·47) 2·14 (1·35–3·39)
Fruit and vegetables 0·69 (0·58–0·81) 0·72 (0·61–0·85) 0·72 (0·59–0.88) 0·77 (0·64–0·93) 0·62 (0·44–0·87) 0·55 (0·38–0·80)
Exercise 0·95 (0·79–1·14) 0·79 (0·66–0·94) 1·02 (0·83–1·25) 0·79 (0·66–0·96) 0·74 (0·49–1·10) 0·75 (0·46–1·22)
Alcohol 1·00 (0·85–1·17) 0·85 (0·73–1·00) 1·03 (0·87–1·23) 0·86 (0·73–1.01) 0·74 (0·41–1·31) 0·83 (0·49–1·42)
All four lifestyle factors 0·20 (0·14–0·27) 0·20† (0·15–0·27) 0·23 (0·16–0·33) 0·21 (0·15–0.29) 0·07 (0·03–0·18) 0·16 (0·06–0·41)
Hypertension 2·24 (1·93–2·60) 1·72 (1·52–1·95) 1·99 (1·66–2·39) 1·72 (1·49–1·98) 2·94 (2·25–3·85) 1·82 (1·39–2·38)
Diabetes 2·96 (2·40–3·64) 2·05* (1·71–2·45) 2·66 (2·04–3·46) 1·93 (1·58–2·37) 3·53 (2·49–5·01) 2·59 (1·78–3·78)
Abdominal obesity 1·79 (1·52–2·09) 1·50 (1·29–1·74) 1·83 (1·52–2·20) 1·54 (1·30–1·83) 1·58 (1·14–2·20) 1·22 (0·88–1·70)
Psychosocial 2·87 (2·19–3·77) 2·43 (1·86–3·18) 2·62 (1·91–3·60) 2·45 (1·82–3·29) 3·92 (2·26–6·79) 2·31 (1·22–4·39)
High ApoB/ApoA1 ratio 4·35 (3·49–5·42) 2·50* (2·05–3·05) 4·16 (3·19–5·42) 2·51 (2·00–3·15) 4·83 (3·19–7·32) 2·48 (1·60–3·83)
All risk factors other than smoking 101·86 (61·22–169·46) 43·24* (26·96–69·37) 59·06 (32·25–108·14) 38·88 (22·95–65·86) 473·43 (158·34–1415·5) 67·49 (21·39–212·90)
All nine risk factors including smoking‡ 216·47 (126·67–369·94) 81·99* (50·02–134·40) 129·19 (68·60–243·28) 76·25 (44·07–131·93) 1100·6§ (342·72–3534·2) 111·45 (32·59–381·12)
Population attributable risks (99% CI)
Lifestyle factors
Smoking 40·7% (35·9 to 45·7) 33·1% (28·9 to 37·6) 52·0% (44·9 to 59·0) 39·0% (34·0 to 44·1) 20·8% (15·7 to 26·9) 8·2% (4·1 to 15·7)
Fruit and vegetables 16·9% (10·8 to 25·3) 11·9% (7·4 to 18·4) 15·7% (8·3 to 27·8) 10·1% (5·3 to 18·2) 18·4% (10·0 to 31·5) 18·7% (10·0 to 32·1)
Exercise 7·5% (0·7·to 46·9) 13·4% (5·4 to 29·7) 0·1% (0·0 to 100·0) 12·5% (4·4 to 30·6) 24·6% (6·8 to 59·2) 23·6% (4·3 to 67·8)
Alcohol –4·1% (–19·8 to 11·6) 11·1% (4·7 to 23·9) –9·1% (–25·1 to 6·9) 10·5% (4·3 to 23·6) 24·9% (3·3 to 76·3) 14·6% (0·5 to 84·6)
All four lifestyle factors 52.1% (39·5 to 64·4) 54·8% (46·2 to 63·1) 55·8% (42·1 to 68·7) 57·1% (48·4 to 65·4) 63·3% (36·8 to 83·6) 51·5% (21·7 to 80·3)
Hypertension 19·2% (16·0 to 22·8) 17·0% (14·0 to 20·5) 12·8% (9·4 to 17·1) 15·7% (12·7 to 19·4) 31·9% (25·7 to 38·6) 25·4% (17·1 to 35·8)
Diabetes 12·4% (10·3 to 14·9) 8·6% (6·9 to 10·7) 8·7% (6·6 to 11·5) 7·8% (6·0 to 10·1) 19·3% (15·1 to 24·5) 13·0% (8·9 to 18·5)
Abdominal obesity 24·8% (17·2 to 34·5) 18·1% (12·2 to 26·0) 23·4% (14·4 to 35·7) 18·3% (11·9 to 27·0) 24·9% (12·4 to 43·7) 11·8% (2·1 to 46·1)
Psychosocial 43·5% (32·2 to 55·6) 25·2% (16·0 to 37·2) 39·7% (25·4 to 56·0) 23·7% (13·9 to 37·4) 53·0% (35·4 to 69·9) 30·6% (10·6 to 62·1)
High ApoB/ApoA·1 ratio 58·9% (50·9 to 66·5) 43·6% (36·6 to 50·8) 59·7% (48·6 to 70·0) 45·3% (37·5 to 53·3) 56·1% (43·7 to 67·7) 36·3% (21·8 to 53·8)
All risk factors other than smoking 89·4% (84·7 to 92·7) 81·7% (76·4 to 86·1) 85·6% (77·7 to 91·0) 80·8% (74·8 to 85·7) 95·5% (90·0 to 98·0) 86·4% (70·8 to 94·3)
All nine risk factors including smoking 93·8% (90·9 to 95·8) 87·9% (84·1 to 90·8) 93·1% (88·9 to 95·8) 88·3% (84·4 to 91·4) 96·5% (92·0 to 98·5) 87·7% (73·1 to 94·9)
*p<0·001 are only provided for the overall comparison. †These values differ slightly but appear similar because of rounding. ‡Based on combining current and former smokers vs never smokers, top two tertiles vs lowest tertile
for abdominal obesity, and top four quintiles vs lowest quintile for ApoB/ApoA1 ratio. If, however, extreme exposures (current vs never, top vs lowest tertile for abdominal obesity, and top vs lowest quintile for ApoB/ApoA1
ratio) were included, the odds ratios for all risk factors for the young group increases to 756·0 and in the old group to 160·8. §Unstable estimate, should be interpreted cautiously.
Table 5: Importance of risk factors in young and old individuals
Region n Control (%) Case (%) Odds ratio (99% CI) PAR (99% CI)
Overall 26527 47·9 65·6 2·27 (2·11–2·44) 36·4% (33·9–39·0)
1403 55·0 73·3 1·96 (1·47–2·62) 29·3% (17·8–44·1)
CE Eur 3624 54·2 69·8 1·92 (1·60–2·30) 30·2% (23·0–38·6)
MEC 3301 45·4 64·5 2·64 (2·19–3·19) 45·5% (39·3–51·9)
Afr 1339 53·8 70·1 2·18 (1·60–2·96) 38·9% (27·1–52·2)
S Asia 3706 41·0 60·3 2·43 (2·03–2·89) 37·4% (31·6–43·6)
China/HK 6062 42·7 62·2 2·30 (2·00–2·65) 35·9% (31·5–40·5)
SE Asia 2131 57·1 70·5 1·96 (1·54–2·49) 36·0% (25·9–47·5)
ANZ 1267 54·2 77·4 2·80 (2·03–3·86) 44·8% (33·0–57·2)
S Am 3068 48·9 69·4 2·35 (1·92–2·87) 38·3% (31·0–46·1)
N Am 626 64·6 76·8 1·82 (1·14–2·88) 26·1% (8·8–56·5)
Odds ratio (99% CI)
0·5 1 2 4 8 160·25
W Eur
Figure 5: Risk of acute myocardial infarction associated with current or former smoking, overall and by region after adjustment for age and sex
W Eur=western Europe. CE Eur=central and eastern Europe. MEC=Middle East Crescent. Afr=Africa. S=South. HK=Hong Kong. SE=southeast. ANZ=Australia and New
Zealand. N=North. Am=America.

Articles
Our data show that risks associated with the major risk
factors (odds ratio of about 2 or greater on univariate
analyses, such as smoking, abnormal lipids,
psychosocial factors, hypertension, diabetes, and
abdominal obesity) were consistently adverse in all
regions of the world and in all ethnic groups. In
particular, the odds ratios for these risk factors were
qualitatively similar (although some quantitative
differences were apparent), despite variations in
prevalence for every risk factor in controls derived from
different subpopulations. However, as expected, the
PAR is affected both by the prevalence of the risk factor
and the odds ratio. We are unaware of any other large
study that has assessed whether risk factors have a
similar or differing effect in many ethnic groups.
Our ﬁnding that most risk factors have directionally
similar odds ratios in ethnic groups and countries differs
from inferences reached by comparison of results of
different studies, which used other methods.3,4
Some of
these researchers suggested that the effects of the major
risk factors could vary qualitatively in different regions
and ethnic groups, possibly because of inconsistent
methodologies, differences in criteria used to recruit
participants, variations in information obtained, and a
fairly modest number of events in each study, thereby
leading to imprecise estimates of risk that could have
been exaggerated or diluted by the play of chance. Since
we had more than 800 cases of acute myocardial
infarction within every major ethnic group (other than
black or coloured Africans), our results within most
Region n Control (%) Case (%) Odds ratio (99% CI) PAR (99% CI)
Overall 21408 20·0 33·5 3·87 (3·39–4·42) 54·1% (49·6–58·6)
W Eur 1047 13·8 29·4 3·76 (2·10–6·74) 44·6% (23·5–67·8)
CE Eur 2618 20·3 29·4 2·20 (1·52–3·18) 35·0% (19·2–54·9)
MEC 3291 29·9 49·6 5·33 (3·48–8·18) 70·5% (57·8–80·7)
Afr 1037 18·0 39·0 7·93 (4·32–14·58) 74·1% (59·7–84·6)
S Asia 2820 27·7 42·7 3·81 (2·49–5·83) 58·7% (42·7–73·1)
China/HK 5400 7·3 14·3 3·43 (2·61–4·51) 43·8% (36·7–51·2)
SE Asia 1858 22·7 48·0 6·22 (3·71–10·41) 67·7% (52·0–80·2)
ANZ 487 13·8 26·8 3·97 (1·71–9·22) 43·4% (16·0–75·6)
S Am 2644 27·1 40·6 2·79 (1·85–4·23) 47·6% (29·6–66·2)
N Am 206 12·4 28·8 4·75 (1·34–16·86) 50·5% (18·2–82·4)
Odds ratio (99% CI)
0·5 1 2 4 8 160·25
Figure 6: Risk of acute myocardial infarction associated with ApoB/ApoA1 ratio (top vs lowest quintile), overall and by region after adjustment for age, sex,
and smoking
PAR is for the top four quintiles versus the lowest quintile.
Region
Overall
W Eur
CE Eur
MEC
Afr
S Asia
China/HK
SE Asia
ANZ
S Am
N Am
Odds ratio (99% CI)
n
26916
1425
3636
3404
1355
3881
6075
2141
1269
3100
630
Control (%)
22·3
16·4
32·7
20·2
21·6
13·8
21·1
15·3
22·0
27·7
28·6
Case (%)
38·6
33·0
46·0
25·9
43·4
31·1
37·3
46·8
37·8
49·3
38·8
Odds ratio (99% CI)
2·48 (2·30–2·68)
2·22 (1·62–3·05)
2·11 (1·76–2·53)
1·84 (1·45–2·33)
3·34 (2·40–4·65)
2·89 (2·31–3·60)
2·41 (2·06–2·80)
5·63 (4·26–7·45)
1·94 (1·40–2·68)
2·48 (2·03–3·04)
1·58(1·01–2·47)
PAR (99% CI)
23·4% (21·7–25·1)
21·9% (15·6–30·0)
24·5% (19·2–30·6)
9·2% (5·7–14·4)
29·6% (22·8–37·3)
19·3% (15·7–23·4)
22·1% (18·7–26·0)
38·4% (33·3–43·6)
22·6% (15·6–31·6)
32·7% (27·2–38·8)
19·0% (9·2–35·2)
0·5 1 2 4 8 160·25
Figure 7: Risk of acute myocardial infarction associated with self-reported hypertension, overall and by region after adjustment for age, sex, and smoking

Articles
ethnic groups are statistically robust. The number of
cases of myocardial infarction in this study within every
region or ethnic group is larger than in most previous
studies, especially in those of non-European origin.
The prevalence of several risk factors varied
substantially, especially when subdivided by sex. For
example, smoking in female controls worldwide has a
prevalence of only 9·25% compared with 33% in male
controls. As a result, despite similar odds ratios in
women and men, the PAR attributable to smoking
varied greatly (16% in women and 44% in men). These
data suggest that the overall approach to prevention of
coronary heart disease could be similar worldwide, but
with varying emphasis in different subgroups (eg, sex
and geographic region) on the basis of the prevalence of
individual risk factors and economic and cultural
factors. The above data also suggest that smoking
cessation is very important in most male populations
worldwide and in women in North and South America,
western Europe, and Australia and New Zealand. By
contrast, quitting smoking is currently less important
for reducing acute myocardial infarction in women in
most other geographic regions. However, if women in
these countries start smoking they are likely to have a
Odds ratio (99% CI)
Region
Overall
W Eur
CE Eur
MEC
Afr
S Asia
China/HK
SE Asia
ANZ
S Am
N Am
n
26903
1422
3636
3401
1355
3882
6075
2140
1269
3093
630
Control (%)
7·6
4·2
6·8
11·6
8·0
10·6
2·9
9·2
4·8
9·0
9·7
Case (%)
18·3
17·7
13·8
25·1
25·0
21·4
11·6
27·4
11·8
20·2
16·6
Odds ratio (99% CI)
3·08 (2·77–3·42)
4·29 (2·61–7·05)
2·61 (1·95–3·49)
2·92 (2·26–3·79)
3·69 (2·38–5·73)
2·48 (1·93–3·18)
5·07 (3·72–6·90)
4·19 (3·01–5·83)
2·37 (1·34–4·19)
2·45 (1·86–3·24)
1·75 (0·94–3·29)
PAR (99% CI)
12·3% (11·2–13·5)
15·0% (11·0–20·1)
9·1% (6·6–12·4)
15·5% (12·2–19·4)
16·7% (11·8–23·1)
11·8% (8·9–15·5)
10·0% (8·3–11·9)
21·0% (16·8–25·9)
7·2% (3·9–12·8)
12·7% (9·4–17·0)
8·0% (2·9–20·1)
0·5 1 2 4 8 160·25
Figure 8: Risk of acute myocardial infarction associated with self-reported diabetes, overall and by region after adjusting for age, sex, and smoking
Region n Control (%) Case(%) Odds ratio (99% CI) PAR (99% CI)
Overall 26136 34·2 47·2 2·24 (2·06–2·45) 33·7% (30·2–37·4)
W Eur 1306 34·1 62·5 4·50 (2·98–6·78) 63·4% (51·3–74·1)
CE Eur 3538 41·0 49·3 1·74 (1·39–2·17) 28·0% (18·3–40·3)
MEC 3309 45·1 58·2 1·85 (1·43–2·41) 25·9% (13·7–43·5)
Afr 1265 38·8 61·1 3·75 ( 2·44–5·74) 58·4% (44·1–71·3)
S Asia 3821 31·2 44·2 2·43 (1·93–3·06) 37·7% (28·8–47·4)
China/HK 6018 21·1 25·3 1·32 (1·11–1·58) 5·5% (1·4–19·4)
SE Asia 2101 21·1 44·4 5·71 (4·13–7·89) 57·9% (49·1–66·2)
ANZ 1224 39·2 63·8 4·40 (2·82–6·88) 61·3% (45·6–74·9)
S Am 2959 48·6 64·8 2·41 (1·79–3·25) 45·5% (32·3–59·2)
N Am 595 31·3 63·7 4·68 (2·57–8·50) 59·5% (40·3–76·2)
Odds ratio (99% CI)
0·5 1 2 4 8 160·25
Figure 9: Risk of acute myocardial infarction associated with abdominal obesity measured as waist/hip ratio (upper tertile vs lowest tertile), overall and by
region after adjusting for age, sex, and smoking
PARs are for top two tertiles vs lowest tertile.

Articles
substantial increase in rates of acute myocardial
infarction attributable to smoking.
Hypertension and diabetes were associated with a
greater odds ratio and PAR in women compared with
men, but women with these factors were about a decade
older than men. Further, the protective effects of exercise
and alcohol consumption also seemed greater in women
than in men. While the amplified effect of diabetes in
women has been reported before,18
we are not aware of
similar data about the other three factors. Thus, even
though significant interactions were noted between these
risk factors and sex for the odds of myocardial infarction, it
would be prudent to seek independent confirmation.
Known risk factors (generally smoking, hypertension,
raised lipids, and diabetes) have sometimes been
claimed to account for only about half the risk of a
myocardial infarction. The origins of this claim are
unclear.19
Our analysis, which is based on traditional and
some newly described risk factors, suggests that more
than 90% of the risk of an acute myocardial infarction in
a population can be predicted by the risk factors
included in our study. Findings of several previous
studies—in which fewer risk factors were measured
(most large studies have not included apolipoproteins,
psychosocial factors or abdominal obesity)—lend
support to our observations. Stamler and colleagues20
studied five US cohorts and categorised individuals on
the basis of the presence of five factors (abnormal
electrocardiogram, diabetes, smoking, cholesterol, and
blood pressure). Those without any of these risk factors
were judged to be in the low-risk category and had an
80–90% lower risk of coronary heart disease in every
cohort compared with the rest of the population. Similar
results were also reported in an analysis of the Göteborg
population, in which individuals with low blood
pressure and a low amount of cholesterol, who were also
non-smokers, had an age-adjusted relative risk of 0·09,
which was much lower that for than the average
population (relative risk 1·0) in the study.21
The importance of modifying risk factors is lent support
by data from randomised trials—eg, blood-pressure
lowering,22
lipid lowering,23
dietary modification24
—or
persuasive evidence of causality from observational
studies25
(eg, smoking cessation).26
Some investigators
have suggested that a pill that combines a statin,
antihypertensive drugs, and aspirin, together with
avoidance of smoking, could potentially reduce the risk of
myocardial infarction by more than 80% to 90%.27
These
studies, along with INTERHEART, suggest that one of the
major emphases in research should be to understand why
currently known risk factors develop in some individuals
and populations, and to identify approaches to prevent
their development or reduce them. For example,
understanding the mechanisms by which societal factors
affect development of risk factors (urbanisation, food and
tobacco policies, shifts in occupation from energy
expending jobs to sedentary ones, and urban structure,
etc) could lead to new approaches to prevent development
of risk factors (primordial prevention),4
which in turn
could reduce coronary heart disease substantially.
Although the odds ratio for an acute myocardial
infarction in people with a family history was about 1·5,
the PAR rose from 90% with the nine potentially
modifiable risk factors to 91% with the addition of family
history. This finding suggests that a large part of the
effect of family history might be mediated through
known risk factors, which could be affected by both
shared lifestyles and genetic factors rather than through
independent pathways. Therefore, the main challenge in
the next few decades will be a combination of
discovering more effective strategies to substantially
alter or prevent development of known risk factors by
understanding the societal, environmental, and
biological causes of the development of these factors.
Odds ratio (99% CI)
Region n Odds ratio (99%CI) PAR (99% CI)
Overall 24767 2·51 (2·15–2·93) 28·8% (22·6–35·8)
W Eur 1375 3·92 (1·95–7·91) 38·9% (20·2–61·5)
CE Eur 3473 1·14 (0·73–1·78) 4·9% (0·0–96·0)
MEC 2892 2·77 (1·78–4·29) 41·6% (25·2–60·1)
Africa 1259 3·17 (1·59–6·33) 40·0% (17·3–67·9)
S Asia 3300 2·15 (1·45–3·17) 15·9% (4·0–46·3)
China/HK 5894 6·00 (4·05–8·89) 35·4% (24·6–47·9)
SE Asia 1921 1·85 (1·02–3·37) 26·7% (17·8–38·1)
ANZ 1255 1·99 (1·02–3·87) 28·9% (10·0–59·7)
S Am 2783 2·08 (1·29–3·36) 35·6% (17·2–59·5)
N Am 615 4·16 (1·50–11·53) 51·4% (24·0–78·0)
0·5 1 2 4 8 160·25
Figure 10: Risk of acute myocardial infarction associated with the composite psychosocial index, overall and by region

Articles
One of the most important risk factors for acute
myocardial infarction in our study was smoking, which
accounts for about 36% of the PAR of acute myocardial
infarction worldwide (and about 44% in men). Regular
consumption of fruits and vegetables was associated
with a 30% relative risk reduction. Thus, eating fruit and
vegetables, taking exercise, and avoiding smoking could
lead to about 80% lower relative risk for myocardial
infarction. Our results are similar to the findings of the
US Nurses Health Study,28
which also indicated that
lifestyle modification could potentially avoid more than
three-quarters of the risks of coronary heart disease and
strokes in women. These conclusions are also lent
support by the results of the Lyon Heart Study,24
which
suggested that dietary modification by itself reduced the
risk of coronary heart disease by about half in patients
with coronary disease. Our data suggest that lifestyle
modification is of substantial importance in both men
and women, at all ages, in individuals from all
geographic regions of the world, and in those belonging
to all major ethnic groups. Therefore, smoking
avoidance, increased consumption of fruits and
vegetables, and moderate activity (along with lipid
lowering) should be the cornerstone of prevention of
coronary heart disease in all populations worldwide.
We also recorded an additional protective effect of
moderate alcohol consumption (PAR 7%). The effect
seemed to be surprisingly large in women, in whom
absence of regular alcohol consumption accounted for
about 22% of PAR, but with wide confidence limits
(–4·9 to 60·8). This finding suggests that the best
estimate of PAR attributable to alcohol consumption in
women is probably closer to the overall estimate of 7%.
Promotion of the consumption of moderate alcohol to
prevent myocardial infarction might also not be
acceptable to many populations, for cultural or religious
reasons, and might increase the proportion of heavy
drinkers and thereby enhance the risk of other diseases
such as strokes, some cancers, cirrhosis of the liver, or
injuries. The overall PAR without alcohol included in the
model is 89·7%; adding alcohol increases it by less than
1% because of the substantial overlap in contributions of
other risk factors. Therefore, advice about alcohol use
could be best customised to individuals depending on
their social, cultural, and religious backgrounds and the
overall effect on their health.
Our study has several potential limitations. First, a case-
control design is potentially open to confounding if there
is differential ascertainment of risk factors between cases
and controls. We minimised this factor by using
standardised methods for data collection in both cases
and controls. The inclusion of incident (first) acute
myocardial infarction cases reduces the possibility that
individuals with previous cardiovascular disease might
have substantially altered lifestyles or risk factor levels
before this event. Further, the odds ratios associated with
all major risk factors—eg, smoking, lipids, diabetes, and
hypertension—in INTERHEART is similar to that
reported in other cohort studies in western populations.
We attempted to minimise biases in the selection of
controls by excluding individuals in whom the risk factors
that we were interested in studying were implicated as
being protective or harmful. Reanalysis of our data by the
two types of controls—hospital-based and community-
based—did not alter our results. Our results are
qualitatively similar for most risk factors in all regions of
the world, providing internal replication. Any selection
biases are unlikely to have been similarly prevalent across
a large number of centres in 52 countries. Therefore, we
think that there is little material bias in our results
because of the use of a case-control study design.
Second, whereas some of the risk factors were
ascertained or measured with high accuracy (eg,
smoking), others (eg, history of diabetes or hypertension)
were based on history and therefore ascertained with
some error. The actual blood pressure value after a
myocardial infarction is potentially confounded because
it might have fallen in some patients because of the
infarction itself or as a result of the drugs used in the
management of the acute phase. Similarly, glucose
concentrations rise with acute myocardial infarction
(stress hyperglycaemia) and are therefore not an
indication of earlier levels. We obtained blood samples
for HbA1c but these are yet to be analysed. Therefore, our
approach to diagnosis of hypertension or diabetes might
have led to misclassification in some individuals with
respect to their risk-factor status. These
misclassifications would tend to underestimate the real
relation between these risk factors and outcomes.
Analysis of our control group data indicates a relation
between the reported prevalence of hypertension in every
centre and measured blood pressure in controls (data not
shown), suggesting that there is some validity in using
self-reports of hypertension as a surrogate for measured
blood pressure. However, the absence of available blood
pressure and glucose values could have underestimated
their importance.
Third, the correlations between repeated measures of
several variables (eg, diet or physical activity) many
months apart is only moderate. Methods to correct for
measurement error and regression dilution bias for one
risk factor have been described;17
however, we are not
aware of methods that adjust for several risk factors
simultaneously. However, if correction for regression-
dilution bias could have been made it could further
increase the odds ratios for most risk factors, which in
turn would increase the overall PAR accounted for by the
nine risk factors that we measured. This outcome means
that the nine risk factors measured in this study probably
account for virtually all the PAR for myocardial infarction
in the population included in this study.
Fourth, our data are based on hospital-based patients
with acute myocardial infarction and matched controls
(mainly from urban areas) and are therefore unlikely to

Articles
reflect the population prevalence of risk factors in an
entire country or region. This fact could potentially have
an effect on our estimates of PAR. However, the key to
ensuring internal validity of the study is to recruit cases
and controls from the same population, which has been
our emphasis. Therefore, our estimates of PAR should be
regarded as providing reliable information about the
specific population enrolled into our study. Nevertheless,
when data are available from several countries (eg, for
smoking), the rates in controls in INTERHEART closely
match published reports for similar age-groups and sexes.
As a result, our overall conclusions that the risk factors
measured in this study account for most of the risk of
acute myocardial infarction is probably broadly applicable.
In view of the consistency of our data, the odds ratios from
the present study could be applied to other populations
and their PAR can then be derived by using population-
specific prevalence rates of specific risk factors.
Fifth, although the effects of individual risk factors and
combinations of four or five of them are reasonably
robust, our estimates of the effect of all nine is model-
dependent because very few individuals have eight or
nine risk factors or, conversely, none. However, crude
examination of the extremes of risk-factor distribution,
and the fact that just five risk factors (smoking, lipids,
hypertension, diabetes, and obesity) for which we have a
sizeable number of individuals predicts about 80% of the
PAR, suggests that our model-based estimates are
reasonably valid.
Our study has several strengths. First, the case-control
study has several advantages over other designs,
especially a cohort study. It allows efficient enrolment of
large numbers of cases and hence greater statistical
power, rapid and cost-effective study conduct, and
enhances the ability to recruit a large number of cases
occurring at young ages, in whom disease association
might be stronger. Second, our study included several
risk factors that have previously not been assessed with
conventional risk factors, including apolipoproteins
(ApoB/A1 ratio), which might be the best marker of the
balance of atherogenic and antiatherogenic particles,10
psychosocial factors,7 and measures of abdominal
obesity, all of which have added substantial information
to the other commonly studied risk factors. Third, the
large size of the study provides high power and precision
in estimates both overall and in subgroups. Fourth, the
inclusion of large numbers of individuals from all
regions of the world and multiple ethnic groups makes
our study results broadly applicable.
In conclusion, our study has shown that nine easily
measured risk factors are associated with more than
90% of the risk of an acute myocardial infarction in this
large global case-control study. These results are
consistent across all geographic regions and ethnic
groups of the world, men and women, and young and
old. Although priorities can differ between geographic
regions because of variations in prevalence of risk
factors and disease and economic circumstances, our
results suggest that approaches to prevention of
coronary artery disease can be based on similar
principles throughout the world. Therefore,
modification of currently known risk factors has the
potential to prevent most premature cases of myocardial
infarction worldwide.
Contributors
S Yusuf initiated the INTERHEART study, supervised its conduct and
data analysis and had primary responsibility for writing this paper.
S Ôunpuu coordinated the worldwide study and reviewed and
commented on drafts. S Hawken did all data analyses and reviewed and
commented on drafts. All other authors coordinated the study in their
respective countries and provided comments on drafts of the
manuscript.
INTERHEART project office staff (Population Health Research Institute,
Hamilton Health Sciences and McMaster University, Hamilton, ON, Canada)
Coordination and Data Management—S Yusuf (principal investigator),
S Ounpuu (project officer), P Montague, J Keys, S Rangarajan (study
coordinators), S Hawken, A Negassa (statisticians), C Wright,
R Mayhew, L Westfall, P Mackie, C Cuvay, D Cunningham, K Picken,
S Anand, H Gerstein.
Core Laboratory—M McQueen, K Hall, J Keys.
INTERHEART national coordinators and investigators
Argentina—R Diaz*, E Paolasso*, M Ciruzzi*, R Nordaby, C H Sosa,
H L Luciardi, E M Marzetti, J A Piasentin, M Amuchastegui, C Cuneo,
G Zapata, J O Bono, E M Marxetti, A Caccavo, R A A Guerrero, R Nordaby,
C Dizeo; Australia and New Zealand—D Hunt*, J Varigos*, M Sallaberger,
G Aroney, P Hicks, D Campbell, J Amerena, W Walsh, G Nelson, D Careless,
J Durech, K Gunawardane, A Thomson, D Owensby, A Buncle, B Watson,
H Ikram, M Audeau, M Hills, D Rosen, J Rankin, A Tunesi, J Sampson,
K Roberts, A Hamer, L Roberts, B Singh, C Singh, R Hendriks, G Tulloch,
A Hill, D Rees, C Hall; Bahrain—M A Halim*, M I Amin; Bangladesh—
S Haque*, G M Faruque, M Hossain; Benin Republic—H Agboton*;
Botswana—C Onen*; Brazil—A Avezum*, L Piegas*, G Reis, J P Esteves,
L C Bodanese, A R Junior, J A M Neto, R F Ramos, E S Silva, A Labrunie,
A C C Carvalho, I P Filho, D C D Albuquerque, J A M Abrantes,
O C E Gebara, P E Leaes, S S Xavier, R L Marino, O Dutra; Cameroon—
W Muna*, K N Blackett*; Canada—K Teo*, S Yusuf*, S, Ounpuu, Y K Chan,
C Joyner, A D Kitching, A Morris, A Panju, H Lee, I Bata, B May,
J G Hiscock, P V Greenwood, W Tymchak, M Natarajan; Chile—F Lanas*,
D Standen, A Lanas, C Santibanez, P Soto, S Potthoff, R Soto, E Mercadal;
China—L S Liu*, L X Jiang*, Q Chen, N Sun, J Wang, L Zhang, X Li,
H Chen, Z Li, Y Shi, J Meng, D Wang, Z Liu, J Yan, W Zhang, Z Wan, Y Xu,
X Li, L Wang, Q Fu, K Zhu, Y Wang, X Wang, F Wang, R Liu, R Zhao, K Sun,
S Yue, Y Zhang, X Zhao, L Zhao, C Zhou, X Meng, A Liu, X Li, R Wu,
R Xiao, X Zhang, G Wan, X Chen, Z Ge, T Zhou, Y Li, X Zhou, F Zhang;
Colombia—L E Bautista*, J P Casas, C A Morillo, E Ardilla, I A Arenas,
M Lindarte, D Aristizabal; Croatia—Z Rumboldt*, M Rumboldt, V Carevic;
Czech Republic—P Widimsky*, M Branny, P Gregor; Egypt—M S Khedr*,
M G Abdel-Aziz, S Abdel-Kader, M A Abdel-Moneim, F El-Demerdash,
A El-Saiid, S Gharib, M Hassanin, F Abdel-Hamid Maklady, I Sadek;
Germany—H J Rupprecht*, T Wittlinger*, P Schuster, A Schmidt, H R Ochs;
Greece—N Karatzas*, A Pipilis*, D Symeonides, K Karidis, T H Tsaknakis,
K Kyfnidis; Guatemala—M Luna*, A M Arroyo, G Sotomora, M A Rodas,
L Velasquez, A Ovando; Hong Kong—J Sanderson*, W Y Ma, S Chan;
Hungary—M Keltai*, F Szaboki, P Karpati, K Toth, S Timar, E Kalo, A Janosi,
M Rusznak, D Lehoczky, J Tarjan, E Kiss, E Sitkei, P Valyi, I Edes; India—
P Pais*, K S Reddy*, P P Joshi, D Prabhakaran, D Xavier, A Roy,
L Ramakrishnan; Iran—M R M Hasani*, S H Mirkhani*; Israel—D Halon*,
B Lewis*; Italy—M G Franzosi*, G Tognoni*, E Gardinale, M Villella,
M Mennuni, F Saiu, G Pettinati, S Ciricugno, S Antonaci, L Carrieri, E Balli,
D Bicego, M G DelleDonne, A Ottaviano, L Moretti, C Melloni, G Melandri,
E Carbonieri, A Vetrano; Japan—M Hori*, H Sato, K Fujii; Kenya—
E N Ogola*, P Wangai; Kuwait—M Zubaid*, W Rashed; Malaysia—
C C Lang*, W A W Ahmad, A Kadirvelu, R Zambahari; Mexico—
M A Ramos-Corrales*; Mozambique—A Damasceno*; Nepal—M R Pandey*,
G Baniya, A Sayami, M Kiratee, B Rawat, M Pandey, J Gurung; Netherlands—
R Peters*, D C G Basart, P N A Bronzwaer, J A Heijmeriks, H R Michels,
G M A Pop; Nigeria—K K Akinroye*; Pakistan—K Kazmi*, N A Memon,
S K Memon, S Nishtar, A Badar, M A Mattu, M Yakub, K Soomra, J Khatri,
A M A Faruqui, S I Rasool, A Samad; Philippines—AL Dans*, F Q Punzalan,
M V C Villarruz, B T Mendoza, C S Recto III, B R Tamesis, V Mendoza,
D J Torres, D D Morales, I Ongtengco; Poland—A Budaj*, L Ceremuzynski*,
S Stec, J Gorny, K Prochniewska, T K Urbanek, P Wojewoda, M Pawlowski,
K Religa, R Klabisz, H Latocha, M Szpajer, K Cymerman, M Laniec,

Articles
H Danielewicz, M Ogorek, D Kopcik, A Baranowska, A Kozlowski,
M Blachowicz, A Jedrzejowski, T Waszyrowski, Z Zielinski, K Janik,
M Tomzynski, M Mytnik, A Maziarz, G Rembelska, M Piepiorka,
K Debkowska, T Krynski, B S Szczeklik, W Krasowski, M Rozwodowska,
P Miekus, J Surwilo, S Malinowski, J Hybel, M Michalak, P Achremczyk,
J Majchrzak, W S Korombel, M Dobrowolski, J Gessek; Portugal—J Morais*;
Qatar—A A Gehani*, H A Hajar*, M M Almowla, A A Omer; Russia—
E G Volkova*, S U Levashov, O M Filatova, U I Evchenko, G S Malkiman,
N N Karaulovskaya; Seychelles—J Panovsky*; Singapore—B A Johan*,
A Cheng, K S Ng, Y L Lim, K H Neoh, K S Tan, L N Lum, Y Cheah; South
Africa—P Commerford*, K Steyn*, B Brown, F Martiz, P Raubenheimer,
A Aboo, J D Marx, E Batiste, K Sliwa, P Sareli, C Zambiakides; Spain—
V Valentin*, J Ferriz, A Rovira, A Rodriguez-Llorian, E Cereijo, T Monzo,
N Alonso, E Gomez-Martinez, A Lopez-Perez; Sri Lanka—S Mendis*,
T Jayalath; Sultanate of Oman—A T A Hinai*, M O Hassan; Sweden—
A Rosengren*; Thailand—C Sitthi-Amorn*, S Chaithiraphan, K Bhuripanyo,
S Srimahachota, M Anukulwutipong, P Loothavorn, P Sritara, S Tanomsup,
J Tantitham, P Tatsanavivat, T Yipinsoi, W Jintapakorn, W Puavilai,
B Koanantakul, P Kasemsuwan, C Supanantareauk, C Peamsomboon,
B Saejueng, K Jeamsomboon, A Sukontasup; United Arab Emirates—
W Almahmeed*, L O Abdelwareth, A Bokhari, N S Rao, S Bakir,
A M Yusufali, E Hatou, Q Zaidi; UK—K Fox*, M Flather*; USA—
J Probstfield*, R Freeman, J Mathew, M Schweiger, W Johnson, E Ofili;
Zimbabwe—J Chifamba*, A J G Hakim.
*National coordinator.
Conflict of interest statement
We declare that we have no conflict of interest.
Acknowledgments
We thank Gilles Dagenais, Sonia Anand, Hertzel Gerstein, and
Janet Wittes for helpful comments on the manuscript; Judy Lindeman
for secretarial assistance; and the staff at the core laboratories in
Hamilton, Canada, and Beijing, China for their excellent efforts.
Further, we thank WHO and the World Heart Federation for their
endorsement, and our friends and colleagues for various forms of help
that led to the successful completion of this global study. SY is
supported by an endowed chair of the Heart and Stroke Foundation of
Ontario and a Senior Scientist Award from the Canadian Institutes of
Health Research (CIHR). SÔ held a Heart and Stroke Foundation of
Canada Fellowship and a Canadian Institutes of Health Research
Senior Research Fellowship during this study. The INTERHEART
study was funded by the Canadian Institutes of Health Research, the
Heart and Stroke Foundation of Ontario, the International Clinical
Epidemiology Network (INCLEN), and through unrestricted grants
from several pharmaceutical companies (with major contributions
from AstraZeneca, Novartis, Hoechst Marion Roussel [now Aventis],
Knoll Pharmaceuticals [now Abbott], Bristol Myers Squibb, King
Pharma, and Sanofi-Sythelabo), and by various national bodies in
different countries (see webappendix2; http://image.thelancet.com/
extras/04art8001webappendix2.pdf).
References
1 Murray CJL, Lopez AD, eds. The global burden of disease: a
comprehensive assessment of mortality and disability from
diseases, injuries, and risk factors in 1990 and projected to 2020.
Boston: Harvard School of Public Health, 1996.
2 Yusuf S, Reddy S, Ôunpuu S, Anand S. Global burden of
cardiovascular diseases, part I: general considerations, the
epidemiologic transition, risk factors, and impact of urbanization.
Circulation 2001; 104: 2746–53.
3 Pais P, Pogue J, Gerstein H, et al. Risk factors for acute myocardial
infarction in Indians: a case-control study. Lancet 1996; 348:
358–63.
4 Yusuf S, Reddy S, Ôunpuu S, Anand S. Global burden of
cardiovascular diseases, part II: variations in cardiovascular disease
by specific ethnic groups and geographic regions and prevention
strategies. Circulation 2001; 104: 2855–64.
5 Ôunpuu S, Negassa A, Yusuf S, for the INTER-HEART
investigators. INTER-HEART: a global study of risk factors for
acute myocardial infarction. Am Heart J 2001; 141: 711–21.
6 Wilhelmsen L, Rosengren A, Johansson S, Lappas G. Coronary
heart disease attack rate, incidence and mortality 1975–1994 in
Göteborg, Sweden. Eur Heart J 1997; 18: 572–81.
7 Rosengren A, Hawken S, Ounpuu S, et al. Association of
psychosocial risk factors with risk of acute myocardial infarction in
11119 cases and 13 648 controls from 52 countries (the
INTERHEART study): case-control study Lancet 2004; 364: 953–62.
8 Marcovina SM, Albers JJ, Kennedy H, et al. International
Federation of Clinical Chemistry Standardization Project for
Measurements of Apolipoproteins A-1 and B: IV comparability of
apolipoprotein B values by use of International Reference Material.
Clin Chem 1994; 40: 586–92.
9 Marcovina SM, Albers JJ, Henderson LO, Hannon WH.
International Federation of Clinical Chemistry Standardization
Project for Measurements of Apolipoprotein A-1 and B: III
comparability of apolipoprotein A-1 values by use of International
Reference Material. Clin Chem 1993; 39: 773–81.
10 Walldius G, Jungner I, Holme I, Aastveit AH, Kolar W, Steiner E.
High apolipoprotein B, low apolipoprotein A-I, and improvement
in the prediction of fatal myocardial infarction (AMORIS study): a
prospective study. Lancet 2001; 358: 2026–33.
11 Breslow N, Day N. Statistical methods in cancer research, vol 1:
the analysis of case-control studies. Lyon: IARC Scientific
Publications, 1980.
12 Walter SD. The distribution of Levin’s measure of attributable risk.
Biometrika 1975; 62: 371–74.
13 Engel LS, Chow WH, Vaughan TL, et al. Population attributable
risks of esophageal and gastric cancers. J Natl Cancer Inst 2003; 95:
1404–13.
14 Bruzzi P, Green SB, Byar DP, et al. Estimating the population
attributable risk for multiple risk factors using case-control data.
Am J Epidemiol 1985; 122: 904–14.
15 Benichou J, Gail MH. Variance calculations and confidence
intervals for estimates of the attributable risk based on logistic
models. Biometrics 1990; 46: 991–1003.
16 Cohen J. A coefficient of agreement for nominal scales.
Educ Psychol Meas 1960; 20: 37–46.
17 MacMahon S, Peto R, Cutler J, et al. Blood pressure, stroke, and
coronary heart disease: part 1, prolonged differences in blood
pressure: prospective observational studies corrected for the
regression dilution bias. Lancet 1990; 335: 765–74.
18 Barrett-Connor E, Cohn BA, Wingard D, Edelstein SL. Why is
diabetes mellitus a stronger risk factor for fatal ischemic heart
disease in women than in men? The Rancho Bernardo Study.
JAMA 1991; 265: 627–31.
19 Canto JG, Iskandrian AE. Major risk factors for cardiovascular
disease: debunking the “only 50%” myth. JAMA 2003; 290:
947–49.
20 Stamler J, Stamler R, Neaton JD, et al. Low risk-factor profile and
long-term cardiovascular and noncardiovascular mortality and life
expectancy: findings for 5 large cohorts of young adult and middle-
aged men and women. JAMA 1999; 282: 2012–18.
21 Rosengren A, Dotevall A, Eriksson H, Wilhelmsen L. Optimal risk
factors in the population: prognosis, prevalence, and secular
trends. Eur Heart J 2001; 22: 136–44.
22 Blood Pressure Lowering Treatment Trialists’ Collaboration.
Effects of different blood-pressure-lowering regimens on major
cardiovascular events: results of prospectively-designed overviews
of randomised trials. Lancet 2003; 362: 1527–35.
23 Heart Protection Study Collaborative Group. MRC/BHF Heart
Protection Study of cholesterol-lowering with simvastatin in 5963
people with diabetes: a randomised placebo-controlled trial. Lancet
2003; 361: 2005–16.
24 de Lorgeril M, Salen P, Martin JL, et al. Mediterranean diet,
traditional risk factors, and the rate of cardiovascular
complications after myocardial infarction: final report of the Lyon
Diet Heart Study. Circulation 1999; 99: 779–85.
25 Parish S, Collins R, Peto R, et al. Cigarette smoking, tar yields, and
non-fatal myocardial infarction: 14000 cases and 32000 controls in
the United Kingdom. BMJ 1995; 311: 471–77.
26 Doll R, Peto R, Boreham J, Sutherland I. Mortality in relation to
smoking: 50 years’ observations on male British doctors. BMJ
2004; 328: 1519–28.
27 Wald NJ, Law MR. A strategy to reduce cardiovascular disease by
more than 80%. BMJ 2003; 326: 1419–23.
28 Stampfer MJ, Hu FB, Manson JE, et al. Primary prevention of
coronary heart disease in women through diet and lifestyle.
N Engl J Med 2000; 343: 16–22.

Interheart risk modifiable factors in micardio infraction 2004

More Related Content

What's hot (18)

Similar to Interheart risk modifiable factors in micardio infraction 2004 (20)

Recently uploaded (20)

Interheart risk modifiable factors in micardio infraction 2004