High levels of serum lipoprotein(a) [Lp(a)] have been associated with increased risk of coronary artery disease (CAD), but this association apparently is not confirmed in elderly people. We evaluated the interactions of Lp(a) with lipid and nonlipid CAD risk factors in a sample of subjects enrolled in the prevalence survey (1992 to 1993) of the Italian Longitudinal Study on Aging (ILSA). The entire population consisted of 5,632 elderly people, aged 65 to 84 years, randomly selected in 8 Italian municipalities. The present cross-sectional study included 400 free-living elderly subjects (74 +/- 6 years) from the randomized cohort of Casamassima (Bari, Southern Italy) (n = 704). The results showed that in the elderly population, high serum Lp(a) is a CAD risk factor dependent on type 2 diabetes mellitus and elevated low-density lipoprotein (LDL) cholesterol levels. In particular, the combined effect of high Lp(a) (> or =20 mg/dl) and high LDL cholesterol (> or =3.63 mmol/L [> or =140 mg/dl]), increases coronary risk by 2.75 (95% confidence interval 7.70 to 0.99); finally, the effect of Lp(a) > or =20 mg/dl and LDL cholesterol > or =3.63 mmol/L (> or =140 mg/dl), combined with type 2 diabetes mellitus, increases risk of CAD by 6.65 (95% confidence interval 35.40 to 1.25). In the elderly, elevated Lp(a) levels appear not to be an independent predictor of CAD, but this lipoprotein is a risk factor only in subjects with type 2 diabetes mellitus and elevated LDL cholesterol.