To guide the design of a nation-wide cohort study of chronic kidney disease in children, we determined how iohexol plasma disappearance curves could be used in children to measure glomerular filtration rate (GFR). Iohexol (5 ml) was administered intravenously and blood samples were obtained at 10, 20, 30, 60, 120, 240, 300, and 360 min after injection (N=29) and assayed by high performance liquid chromatography. Four urines were also collected following the injection. Intra-assay coefficient of variation (CV) in serum was 1.3% at 100 mg/l, 2.6% at 15 mg/l, and 3.4% for duplicate unknowns. GFR(9) was computed from iohexol dose and area under the nine-point blood disappearance curve, using double exponential modeling. Only 2.8% of 254 data points deviated by >3 CV from the curves. GFR(4) calculated from 10, 30, 120, and 300 min points correlated well with GFR(9) (r=0.999) and showed no bias (means+/-s.d. of GFR(9) and GFR(4)=59.3+/-36.3 and 59.4+/-36.0 ml/min per 1.73 m(2)). Relationship of GFR(9) and one-compartment GFR followed quadratic equation as previously reported by Brochner-Mortensen, allowing GFR to be calculated from 120 and 300 min points. This GFR(2) correlated well with GFR(9) (r=0.986). Estimated GFR from Schwartz height/creatinine formula correlated with GFR(9)(r=0.934) but overestimated GFR by 12.2 ml/min per 1.73 m(2). Urine iohexol clearance was poorly correlated (r=0.770) with GFR(9) owing to variability in urine collections (median CV=24%). GFR can be measured accurately using four-point iohexol plasma disappearance (in most cases, two points suffice); estimated GFR and urinary clearances are less useful.