1. Recent outcome trials suggest that lipid-lowering with 3-hydroxy-3-methylglutaryl-coenzyme A reductase inhibitors is justifiable on risk-benefit grounds in subjects with serum cholesterol > 5.5 mmol/l who have coronary heart disease, other forms of atherosclerotic vascular disease, or who are free of vascular disease but have a risk of major coronary events > or = 1.5% per year. Choice of an appropriate treatment policy will require (i) knowledge of the proportion of the population who will need treatment for secondary prevention, and (ii) targeting of treatment for primary prevention at a specified absolute risk of coronary heart disease events. Selection of an appropriate coronary heart disease risk for primary prevention requires consideration of the number needed to be treated to prevent one coronary heart disease event, the proportion of the population requiring treatment, the cost-effectiveness of treatment and the total cost of treatment. 2. In a random stratified sample of subjects aged 35-69 years from the Health Survey for England 1993 we first examined the prevalence of subjects with cardiovascular disease and serum cholesterol > 5.5 mmol/l who may be candidates for secondary prevention. In those free of cardiovascular disease we then examined the prevalence of subjects with serum cholesterol > 5.5 mmol/l who had three different levels of coronary heart disease risk: coronary heart disease event rates of 4.5% per year, 3.0% per year and 1.5% per year. These subjects may be candidates for primary prevention depending on the treatment policy selected. 3. For secondary prevention, 4.8% (95% confidence interval 4.3-5.3) of the U.K. population aged 35-69 years might be candidates for 3-hydroxy-3-methylglutaryl-coenzyme A reductase inhibitor treatment, comprising 2.4% (2.0 to 2.7) with a history of myocardial infarction, 1.9% (1.6 to 2.2) with angina and 0.5% (0.3-0.7) with a history of stroke--all with total cholesterol > 5.5 mmol/l. The prevalence of these diagnoses with total cholesterol > 5.5 mmol/l increased with age, from 1.5% at age 35-39 years to 16.2% at age 65-69 years in men, and from 0.2% at age 35-39 years to 10.0% at age 65-69 years in women. Approximately 13 people would need treatment for 5 years to prevent one coronary event, at a cost of 36,000 pounds per event prevented. The number needing treatment for secondary prevention would increase substantially if treatment was extended to patients above 70 years of age or to those with serum cholesterol < or = 5.5 mmol/l. 4. Primary prevention aimed at a coronary event risk of 4.5% per year would lead to treatment of only 0.3% (0.2-0.4) of those aged 35-69 years, and those treated would be predominantly older men with additional risk factors for coronary heart disease. The number needed to be treated and cost per coronary event prevented would be similar to those for secondary prevention. 5. Primary prevention targeted at subjects with a coronary event rate of 3.0% per year would entail treating 3.4% (3.0-3.9) of all those aged 35-69 years. At this level of risk, 20 people would need treatment for 5 years to prevent one coronary event, at a cost of 55,000 pounds per event prevented. 6. Primary prevention aimed at a coronary event rate of 1.5% per year would entail treating 19.6% (18.7-20.6) of all subjects aged 35-69 years, and about 80% of men aged 60-69 years for primary or secondary prevention. At this level of risk, 40 people would need treatment for 5 years to prevent one event, at a cost of 111,000 pounds per event saved. 7. Guidelines for 3-hydroxy-3-methylglutaryl-coenzyme A reductase inhibitor treatment should take into account the considerable workload and financial resources needed to implement secondary prevention of coronary heart disease, the accepted first priority. For primary prevention they need to consider the number needed to be treated to prevent one event, the number of subjects needing treatment, the cost-effectiveness of treatment and