PURPOSE: This study assesses the usefulness of alkaline phosphatase (ALP) in predicting the radiologic osteopenia and rickets for very low birth weight infants (VLBWI) below 1,500 g of birth weight. METHODS: Medical records of 258 VLBWI who were admitted at NICU between January 2009 and December 2011 were reviewed retrospectively. ALP max was defined by maximum level of ALP during admission. Subjects were divided into two groups: the study (n=112, ALP max > or =450 IU/L) and the control group (n=146, ALP max <450 IU/L). The two groups were compared based on the demographic and perinatal risk factors, clinical parameters, and morbidity. Also compared between the groups were: the level of ALP max, total and ionized calcium, inorganic phosphorus, and radiologic finding. To assess the usefulness of ALP in predicting the radiologic osteopenia and rickets, the sensitivity, specificity, and the positive (PPV) and negative predictive values (NPV) according to ALP max level were examined. A receiver operator characteristic (ROC) curve was constructed to determine the optimal cut-off value of ALP for the radiologic diagnosis. RESULTS: The study group was more immature and smaller, and had more perinatal risk factors with higher morbidity. The sensitivity and specificity for the radiologic finding on the basis of ALP max 450 IU/L were 68.8% and 64.9%, PPV 39.3%, and NPV 86.3% with a significance (P<0.001). In ROC curve, we observed the appropriate cut-off value to be 459.5 IU/L (P<0.001). CONCLUSION: The baseline of ALP max 450 IU/L was useful in predicting the radiologic osteopenia and rickets in VLBWI.