To assess the ability of childhood blood pressure, height, and weight to predict young adult blood pressure, the authors examined data obtained over multiple visits for four years on 339 children aged 8-18 years in East Boston, Massachusetts. These subjects were again seen 8-12 years later when they were aged 20-26 years. Multivariate regression models were used to predict true blood pressure in young adulthood from observed childhood measurements closest to age 10 (n = 219), adjusting for within-person variability. Without adjusting for childhood blood pressure, childhood height, weight and body mass index were at least marginally associated with young adult systolic blood pressure in boys and girls, with similar coefficients for each gender. The strongest predictor was weight (beta = 0.6 mmHg/10 lbs for girls, and beta = 0.7 mmHg/10 lbs for boys), and height was no longer predictive with weight in the model. With childhood blood pressure included, neither childhood height nor weight were predictors of future systolic blood pressure. However, change in height and weight were predictors of future systolic blood pressure. Weight change was a stronger predictor in girls than boys with beta = 0.9 mmHg/10 lbs. For diastolic blood pressure, height and weight had limited predictive ability in these data. These models, which allow for both between- and within-person variability in young adulthood, may be used to estimate the predictive value for future high blood pressure of a child's current blood pressure, height and weight, as well as future change in height and weight. These data suggest that the effects of childhood height and weight on future blood pressure may be negligible given childhood blood pressure, but that later height and weight remain predictive.