PURPOSE: There have been high rates of false positive and recall in neonatal screening test using the cut-off points set by the manufacturing company. So, it is necessary to re-evaluate the cut-off values to minimize the false positive rates. METHODS: We collected capillary blood in dry filter paper from 996 healthy neonates on the third day in cases of normal vaginal delivery or the fifth day in cases of Cesarean section. The levels of phenylalanine, galactose, 17-hydroxyprogesterone and branched-chain amino acids were measured using enzyme immunoassay. The results were compared with the original cut-off points set by the manufacturing company. RESULTS: The original cut-off points of four substances were 4.0mg/dL, 7.5mg/dL, 35ng/mL, and 8.0mg/dL, respectively, so that false-positive rates were 0.4, 1.6, 3.93, and 0.001%, respectively. When we set the cut-off point at 99.7 percentile using the data from healthy neonates, they should be 4.0mg/dL, 9.2mg/dL, 54.3ng/mL, and 8.0mg/dL, respectively. CONCLUSION: The false-positive and recall rates were higher in galactosemia and congenital adrenal hyperplasia when using the original cut-off points, suggesting that it would be reasonable to modify the cut-off point at 99.7 percentile after measuring those substances from enough of healthy neonates.