Abstract

Characteristics of 11,132 men aged 40-59 years and free from coronary heart disease (CHD) at entry were related to follow-up experience, using multivariate analysis. In 5 years among 2,404 U. S. railroad men and 8,728 European men there were 615 cases of CHD, 214 of whom died from CHD or suffered definite nonfatal infarction. With five entry characteristics (age, systolic blood pressure, serum cholesterol, smoking habit, and body mass index), multiple logistic solutions for Europeans and Americans, separately, gave estimates of the individual probability of CHD. Classified by deciled scores for these probabilities, the expected and observed CHD cases were highly correlated (r = 0.930-0.981). Predictions based on European data applied to Americans, and vice versa, gave similar high correlations but American incidence was excessive compared with European experience. Application of the analysis coefficients obtained from data in Europe and in the U. S. railroad to 6,221 other U. S. men 40-59 years of age, CHD-free at entry, gave good prediction of relative risk (r = 0.94) for observed versus predicted cases in deciles of risk score; however, the actual numbers of cases were underpredicted. From single measurements of a few characteristics the multiple logistic solution usefully estimates the relative risk of CHD for individuals. Age, systolic pressure, and serum cholesterol are universally powerful predictors of risk. Variables not measured in this study or not yet identified contribute to the risk of CHD among American men.