We investigated whether serum concentrations of apolipoprotein (apo) B and apoA-I and the apoB/apoA-I ratio provided predictive information on myocardial infarction (MI) and ischemic stroke during 13 years of follow-up in a group of initially clinically healthy 58-year-old men, free from previous cardiovascular disease, diabetes, other established disease, or treatment with cardiovascular drugs. Multivariate logistic regression analysis showed that the apoB/apoA-I ratio and apoB were significant and independent determinants of MI (exponentiation of the B coefficient [Exp(β)] 3.1, 95% confidence interval [CI] 1.6-6.3, P = .001, Exp(β) 2.8, 95% CI 1.1-7.7, P = .045, respectively). The area under the receiver–operating characteristics curve as a relative measure of test efficiency was highest and significant for both apoB/apoA-I ratio and apoB (area under the curve = 0.75, P < .001). In conclusion, the apoB/apoA-I ratio and apoB are independent risk factors for MI and has the highest efficiencies for predicting MI in initially healthy middle-aged men.