It has been standard of care in our NICU to obtain a screening (HUS) on infants born at <34wks gestational age (GA) or <1500g birthweight (BW). Our goal was to develop new screening criteria in order to: 1) identify only those infants with IVH associated with adverse neurodevelopmental outcome(grade III-IV); and 2) reduce the costs associated with unnecessary screening HUS. We conducted a retrospective chart review to identify the clinical indicators in the first 7 days of life that were most predictive of grade III-IV IVH. All infants born at <34wks GA from 1/1/93-12/31/93 who had at least one HUS were included in the analysis (n=188). Infants with grade III-IV IVH (n=16) were then compared with the infants with no IVH or grade I-II IVH(n=172) by student t-test or chi square analysis and the results were expressed as mean ± SD or as a percent. There was no difference between the two groups with regard to sex, 5 min Apgar <5, multiple birth, seizures, anemia defined as Hct<30, pneumothorax, or treatment with antenatal betamethasone. When compared with the infants with clinically insignificant IVH, the infants with grade III-IV IVH had a lower BW (991g± 443 vs 1409g ± 445; p<0.001), lower GA (27.0wks ± 3 vs 30.0wks ± 3; p<0.001), were more likely to require mechanical ventilation and surfactant (68.8% vs 32.6%; p=0.027), had base deficit ≥10(50% vs 15.7%; p=0.003), thrombocytopenia with platelet count <50,000(18.8% vs 2.9%; p=0.022), hypocapnia defined as PaCO2 <30 (80.0% vs 28.4%; p<0.001), hypotension requiring fluid bolus and/or pressor support(87.5% vs 29.1%; p<0.001), or PDA requiring treatment (68.8% vs 29.8%; p=0.004). When adjusted for BW, the risk of grade III-IV IVH was significantly increased among infants with base deficit, thrombocytopenia, hypocapnia, or hypotension. Logistic regression analysis performed using the significant clinical variables demonstrated that the risk of grade III-IV IVH was increased among those infants with hypotension (OR:2.62; 95%CI:1.15-5.98) or hypocapnia (OR:2.22; 95%CI:1.12-4.44). BW or GA were not significant predictors of risk. These data suggest that screening HUS should be performed according to clinical indicators and not necessarily BW or GA.