Abstract
OBJECTIVE: To investigate the hypothesis that the apparent protective effect of habitual alcohol consumption on coronary heart disease is due to drinkers at high risk of coronary heart disease becoming non-drinkers. DESIGN: Case-control population based study. Data were obtained from interviews with patients with non-fatal myocardial infarction and their controls and with the next of kin of those who had died of coronary heart disease and their controls. SETTING: Auckland, New Zealand. SUBJECTS: Two groups of cases were studied. The first comprised 227 men and 72 women with non-fatal myocardial infarction identified from a population based surveillance programme for coronary heart disease; controls were 525 men and 341 women randomly selected from the same population group and matched for age and sex. The second group comprised 128 men and 30 women who had died of coronary heart disease and had been identified from the surveillance programme; controls were a sample of the previous control group and comprised 330 men and 214 women matched for age and sex. All participants were aged 25-64 years and without diagnosed coronary heart disease. MAIN OUTCOME MEASURES: Regular alcohol consumption; high density lipoprotein cholesterol and low density lipoprotein concentrations. RESULTS: Men with myocardial infarction and men who had died of coronary heart disease were more likely to have been never drinkers (had never drunk more than once a month) than controls (18% v 12% and 23% v 13% respectively). After possible confounding factors had been controlled for, people in all categories of drinking (up to more than 56 drinks per week) had at least a 40% reduction in risk of fatal and non-fatal coronary heart disease compared with never drinkers. Former drinkers also had a lower risk of non-fatal myocardial infarction than never drinkers (relative risks 0.41 and 0.10 in men and women respectively) but a similar risk of death from coronary heart disease. The reduction in risk was consistently greater in women than in men in all drinking categories but there was no clear dose-response effect in either sex. CONCLUSIONS: The results support the hypothesis that light and moderate alcohol consumption reduces the risk of coronary heart disease. This protective effect in this population was not due to the misclassification of former drinkers with a high risk of coronary heart disease as non-drinkers.