Objective-Carotid artery intima-media thickness (IMT) is an intermediate phenotype for atherosclerosis. In a community population (nϭ5400), we determined whether a parental history of myocardial infarction (MI) or stroke is associated with increased IMT and whether associations differ at specific sites in the carotid arterial tree. Methods and Results-Using regression modeling, the proportion of IMT that remains unexplained after controlling for vascular risk factors was determined. A parental history of stroke was associated with both increased common carotid artery (CCA) and increased internal carotid artery (ICA)-IMT, but in young individuals (Յ60 years of age), the association was stronger with ICA-IMT, with an odds ratio (95% CI) for ICA-IMT in the highest quartile of 2. been widely used to study genetic and other novel risk factors for atherosclerosis is carotid artery intima-media thickness (IMT). This can be estimated noninvasively using ultrasound and seems to reflect subclinical atherosclerosis. Crosssectional studies have shown that increased common carotid IMT is both a marker of atherosclerosis elsewhere in the arterial system 3 and an independent predictor of future stroke and myocardial infarction (MI) risk. 4 -6 Carotid IMT measurements are increasingly being used as an intermediate phenotype for genetic studies of vascular disease. 7-9 Their use has several advantages. First, the number of genes involved in this intermediate phenotype are likely to be considerably less than those involved in ischemic stroke or MI. Second, the technique can be readily applied to large-scale community-based populations, reducing selection bias. Third, the use of a continuous index of risk rather than a dichotomous variable such as stroke and the related avoidance of nonpenetrance of asymptomatic disease markedly increase power.IMT measurements are strongly influenced by conventional risk factors. 10 This may complicate and reduce the power of studies to investigate genes that independently increase IMT rather than those that merely predispose to conventional risk factors. Spence et al 11 devised a novel method of combining quantitative measures of carotid plaque area measured by ultrasound and linear regression modeling to determine unexplained atherosclerosis, ie, the proportion of atherosclerosis that remains unexplained after controlling for conventional vascular risk factors. This approach was then used to identify subjects with novel risk factors for atherosclerosis, in this case, elevated homocysteine. 11 A drawback of this method is that measurements can only be obtained in subjects with established carotid plaque, and it is therefore not applicable to a general community population. We have applied this method to carotid IMT measurements instead of plaque area, using similar regression modeling, to determine the extent of IMT that remains unexplained by conventional