Objective: To assess the accuracy of specific plasma lipid fractions in predicting coronary heart disease (CHD) mortality among adults.
Methods: Follow-up data for a random sample of 30- to 79-year-old men and women recruited into the Lipid Research Clinics Prevalence and Follow-up Studies were included in this analysis (n = 4499). Baseline measurements of total plasma cholesterol and lipoprotein fractions were compared with subsequent CHD mortality after a mean follow-up of 12.3 years. The areas under receiver operating characteristics curves for specific serum lipids were compared for individuals aged 30 to 59 and 60 to 79 years.
Main results: For the younger cohort, the ratio of total cholesterol to high-density lipoprotein cholesterol was a better predictor (P < .05) of CHD mortality (receiver operating characteristic curve area, 0.80 +/- 0.03) than was total cholesterol level alone (receiver operating characteristic curve area, 0.73 +/- 0.03) or any other single lipoprotein measurement. Among the older cohort, the same screening strategies performed poorly, with receiver operating characteristic curve areas ranging from 0.51 +/- 0.05 for total cholesterol to 0.64 +/- 0.05 for the ratio of low-density to high-density lipoprotein cholesterol levels.
Conclusion: Plasma lipid levels are poor predictors of coronary death among those aged 60 to 79 years without known CHD. These data indicate the need to define better lipid screening strategies for older, asymptomatic adults. Among younger adults aged 30 to 59 years, high-density lipoprotein cholesterol measurement should be included as part of any lipid screening program, as the ratio of total to high-density lipoprotein cholesterol levels is the best lipid screening test to identify those at high risk for subsequent CHD mortality.