The use of the onset of the fourth (K4) or fifth (K5) Korotkoff phase to determine diastolic blood pressure in children has been controversial; most recently, the Second Task Force recommended the use of K4 for children up to age 13 years and K5 for children age 13 and above. We performed a cross-sectional analysis of 1,155 nine-year old (53% white and 47% black) and 1,224 ten-year old girls (45% white and 55% black) in the National Heart, Lung, and Blood Institute Growth and Health Study (NGHS). The mean blood pressures for the first, fourth, and fifth Korotkoff phases were 100.1 (+/- 8.9) mm Hg, 66.6 (+/- 9.8) mm Hg, and 56.8 (+/- 11.8) mm Hg for nine-year-olds and 102.8 (+/- 9.0) mm Hg, 68.1 (+/- 10.1) mm Hg, and 58.1 (+/- 11.9) mm Hg for ten-year-olds. The mean difference between K4 and K5 was 9.9 (+/- 6.4) mm Hg. The correlation between K1 and K4 was 0.45, between K1 and K5 was 0.34, and between K4 and K5 was 0.84. Elevation of blood pressure was defined at or above the 95th percentile based on the NGHS distribution for K1, K4, or K5; the relative risk of having an elevated K1 was 10.1 if K4 was elevated and 5.9 if K5 was elevated. Of the 159 subjects potentially classified with elevated diastolic pressure, 95 subjects (60%) would be classified differently depending on whether K4 or K5 was used to define elevated diastolic blood pressure. The choice of the onset of the fourth or fifth Korotkoff phase for determining diastolic blood pressure in children may have important implications for which individuals are classified as having hypertension.