Table 2 Median age (years) of presentation of cases
Related Figures (22)
Table 1: Distribution of study population 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. tTop 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 lowes quintile for ApoB/ApoA1, the odds ratio for the combined risk factors increases to 333-7 (99% Cl 230-2-483-9). 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). 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 Figure 1: Odds of myocardial infarction according to number of cigarettes smoked and ApoB/ApoA1 ratic Note the doubling scale on the y axis for both figures. 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). 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. 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% Cl 90:24-184-99). Risk by age 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 model. 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% Cls is shown in webtable 3 (http://image.thelancet.com/extras/04art8001webtable3.pdf). *Saturated model, no difference between adjusted and unadjusted models. {Non-estimatable. *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 tert for abdominal obesity, and top four quintiles vs lowest quintile for ApoB/ApoA11 ratio. If, however, extreme exposures (current vs never, top vs lowest tertile for abdominal obesity, and top vs lowest quintile for ApoB/ApoA: 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. 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. 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. 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 Figure 6: Risk of acute myocardial infarction associated with ApoB/ApoA 1 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. Figure 7: Risk of acute myocardial infarction associated with self-reported hypertension, overall and by region after adjustment for age, sex, and smoking 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.** 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 Figure 8: Risk of acute myocardial infarction associated with self-reported diabetes, overall and by region after adjusting for age, sex, and smoking 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. 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 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 Figure 10: Risk of acute myocardial infarction associated with the composite psychosocial index, overall and by region *Matched models included additional adjustment for smoking status for all models where smoking was not the risk factor of focus. {Unmatched models included adjustment for age, sex, region, and smoking (where smoking was not the risk factor of focus). The matched analysis data subset was used in the unmatched analyses for direct comparability.
