As the correlation of bone mass from childhood to adulthood is unclear, we conducted a long-term prospective observational study to determine if a pediatric bone mass scan could predict adult bone mass. We measured cortical bone mineral content (BMC [g]), bone mineral density (BMD [g/cm(2)]), and bone width (cm) in the distal forearm by single photon absorptiometry in 120 boys and 94 girls with a mean age of 10 years (range 3-17) and mean 28 years (range 25-29) later. We calculated individual and age-specific bone mass Z scores, using the control cohort included at baseline as reference, and evaluated correlations between the two measurements with Pearson's correlation coefficient. Individual Z scores were also stratified in quartiles to register movements between quartiles from growth to adulthood. BMD Z scores in childhood and adulthood correlated in both boys (r = 0.35, p < 0.0001) and girls (r = 0.50, p < 0.0001) and in both children ≥10 years at baseline (boys r = 0.43 and girls r = 0.58, both p < 0.0001) and children <10 years at baseline (boys r = 0.26 and girls r = 0.40, both p < 0.05). Of the children in the lowest quartile of BMD, 58 % had left the lowest quartile in adulthood. A pediatric bone scan with a value in the lowest quartile had a sensitivity of 48 % (95 % confidence interval [CI] 27-69 %) and a specificity of 76 % (95 % CI 66-84 %) to identify individuals who would remain in the lowest quartile also in adulthood. Childhood forearm BMD explained 12 % of the variance in adult BMD in men and 25 % in women. A pediatric distal forearm BMD scan has poor ability to predict adult bone mass.