Patent ductus arteriosus (PDA) is one of the most common complications in premature infants. Successful pharmacological closure of PDA with indomethacin was first reported in 1976. Since then indomethacin treatment has become the standard or prophylactic treatment for clinically significant PDA in premature infants. Clinically there is a high incidence of complications associated with indomethacin treatment, including hypoglycemia, necrotizing enterocolitis, GI bleeding, extension of IVH. More recently, ibuprofen has been shown to be effective for the closure of patent ductus arteriosus in premature infants without reducing mesenteric, renal, or cerebral blood flow.Ibuprofen has been shown to close the ductus in animals without reducing cerebral,intestinal or renal blood flow. Furthermore, ibuprofen enhanced cerebral blood-flow autoregulation and had some neuroprotective effect. In recent years, our strategy of PDA treatment for ELBW infants was essentially early targeted indomethacine treatment depending on echocardiographic shunt flow pattern of PDA. (Arch Dis Child 1997;77:F36-F40. Acta Paediatr Tw 1998;39:33-7. and Arch Dis Child 1999;79: F197-F200.) By this regimen, infants will be eligible for the study if their birth weight less than 1000 gm and if they had PDA without other structured cardiac anomaly confirmed by echocardiography shortly after birth (as close as possible to12 hours). After parental informed consent is obtained, infants will be randomly assigned to two groups based on a double-blined design. INDO group will receive echocardiographic assessment at interval of 12-24 hours or clinically necessary, and if the PDA had pulsatile or growing flow pattern, indomethacin is given; if the PDA had flow patterns other than growing or pulsatile pattern, no treatment is given. The subsequent dose of indomethacin is according to the echocardiographic flow patterns at interval of 24 hours from the last dose. When indomethacin was fail to close after the first course, the second course of another 3 doses of indomethacin or ibuprofen will be given. In spite of infants of INDO group or IBUO group, if PDA fail to close after 2 courses of treatment, surgical ligation of PDA would be considered according to the infant’s clinical condition. Our historical data showed that the incidence of complication was about 30%. Permitting 5% chance of type I error and 20% of type II error and an absolute reduction of the incidence by 20%, 30 infants in each group is needed to detect a difference. Primary outcome of the assessment is the closure rate of PDA and the incidence of death or pulmonary hemorrhage. Secondary outcome is IVH or PVL, NEC, oliguria and CLD. We expect that, by using this treatment regimen, a high PDA closure rate can be achieved and the survival of very premature infants may be increased.