That cardiovascular (CV) risk variable levels tend to track in the same ranking over time is well recognized. However, information on tracking of CV risk variables over a long period of time from childhood to adulthood and the changes in tracking with increasing intervals is limited. This aspect was examined in a biracial (black-white) cohort of 7770 individuals (63% white, 46% male) who were examined repeatedly 2-12 times with 31500 observations over an average follow-up period of 12 years (range: 1-28 years). Pearson correlations of pairs of two consecutive values measured 1 to 28 years apart decreased progressively regardless of risk variables as the intervals between the two observations increased. Intraclass correlation, an overall measure of tracking, was estimated using 4 serial measurements over 20-25 years, 2 in childhood (4-17 years), and 2 in adulthood (18-43 years). Body mass index showed the greatest intraclass correlation (0.74 for whites; 0.73 for blacks), followed, in order, by low-density lipoprotein cholesterol (0.56; 0.58), systolic blood pressure (0.43; 0.38), high-density lipoprotein cholesterol (0.35; 0.36), diastolic blood pressure (0.30; 0.34), and triglycerides (0.36; 0.28). Both the Pearson and intraclass correlations of all risk variables did not differ by race and sex groups and remained essentially the same when adjusted for body mass index. Thus, CV risk variables measured in childhood are predictive of adulthood levels over a period of 2-3 decades. These results underscore the value of measuring childhood CV risk variables in identification of high-risk individuals, which is important for early prevention and intervention.