Objective: To investigate the relative roles of serum apolipoprotein (apo) B and low density lipoprotein (LDL)-cholesterol levels in predicting incident coronary heart disease (CHD). Whether apo B/apo A-I ratio has advantage over apo B in this prediction constitutes a secondary aim. Methods: Prospective evaluation of 1138 men and 1210 women, aged 28-74 years participating in the TEKHARF survey 1997/98 with a mean 5.9-years’ follow-up in whom serum apo B was determined. Tertiles of LDL-cholesterol were formed by cut points of 130 and 100 mg/dl, and of apo B by 120 and 95 mg/dl. Metabolic syndrome was defined by modified ATPIII criteria. Nonfatal CHD diagnosis was based on history of angina and myocardial revascularization, physical examination of the cardiovascular system and Minnesota coding of resting electrocardiograms. Results: Apolipoprotein B showed significant correlations with a greater number of parameters than did LDL-cholesterol. Incident CHD was not significantly predicted in age-adjusted logistic regression by LDL-cholesterol but by apo B concentrations in men with a relative risk (RR) 1.005. Apolipoprotein B level >120 vs <95 mg/dl retained significance in both genders combined, even after adjustment for waist girth and log C-reactive protein. The top (>1.02) compared with the bottom bracket of apo B/A-I ratio, though not reaching significantly predictive values among women, did significantly predict in men incident CHD with a RR 1.89. Conclusions: Apolipoprotein B, which marks small, dense LDL particles in plasma is a better predictor of incident CHD than LDL-cholesterol among Turkish adults. While in the prediction of CHD apo B level should be preferred in women, an apo B/A-I ratio >1.02 has advantages over the latter in men. It is time to create the environment (at least in our cities) for measuring in equipped laboratories apo B, which has advantages over measurements that permit calculation of LDL-cholesterol.