In a defined population of 4349 men and 3398 women, we evaluated the frequency and clustering of five nonlipid coronary heart disease risk factors--obesity, hypertension, sedentary lifestyle, cigarette smoking, and alcohol consumption--in nine categorical lipoprotein phenotypes--normal, types I or V, IIA, IIB, III, IV, hypoHDL (high-density lipoprotein), hypoLDL (low-density lipoprotein), and hyperHDL. This analysis indicated that compared with the normal phenotype, nonlipid coronary risk factors tend to be more frequent and thus cluster in individuals with "high risk" phenotypes, and occur somewhat less frequently in those with "low risk" phenotypes. Thus, identification of a high-risk phenotype suggests the presence of nonlipid risk factors as well, and the clinician should be alert to this possibility. A multivariable analysis of the independent associations of each of the risk factors with the lipid and lipoprotein components that define the phenotypes suggested that several behavioral risk factors may directly affect lipid and lipoprotein levels. This observation raises the possibility that certain intervention strategies, such as weight loss, smoking cessation, and regular exercise, may favorably influence dyslipoproteinemia.