In a prospective cardiovascular study of 1341 Trinidadian men aged 35-69 years undertaken between 1977 and 1986, the baseline prevalence rates of cardiac and arterial disease and diabetes mellitus were increased in the 118 (8.8%) who had been but were no longer regular drinkers. This finding suggested that awareness of these disorders was a discouragement to drinking alcohol. When this group and all with coronary heart disease (CHD) or diabetes at entry were excluded, a significant inverse trend was found between alcohol consumption in the week before recruitment and risk of CHD across the subsequent average follow-up of 7.5 years. Men who had taken 5-14 drinks had about half the CHD risk of those who had had no alcohol, even after allowance for age, ethnicity, smoking, blood pressure and serum cholesterol concentration. The overall morbidity and mortality experience in this community indicated a protective effect of alcohol against CHD, but adverse health consequences from multiple causes in drinkers who were alcohol dependent.