Data from a population based cohort study [Pinto-Martin et al. Pediatrics 1995] suggests that almost one-half of VLBW infants with CP do not have antecedent major CUS abnormalities, such as periventricular echodensity or echolucency (PVED/EL). We performed a case-control study to identify risk factors for CP among VLBW infants without major CUS abnormality. 36 cases, who had CP and no major CUS abnormality, and 70 gestational-age matched controls were identified within a geographically based cohort, born 1985 to 1995, at the only tertiary perinatal referral center serving a 17-county region. 27 cases had normal CUS studies and 9 had uncomplicated subependymal hemorrhage(SEH), defined as SEH without post-hemorrhagic hydrocephalus, persistent ventricular enlargement, or PVED/EL. Medical records of mothers and neonates were reviewed to ascertain the occurrence of the following factors, which, in previous studies, have been associated with either CP or PVED/EL: multiple gestation, chorioamnionitis, pre-eclampsia, treatment of the mother with magnesium sulfate or betamethasone, neonatal sepsis, pneumothorax, acidosis, hypotension, hypocarbia, and bronchopulmonary dysplasia (BPD). In univariate analyses, betamethasone was associated with a lower risk [odds ratio (OR) 0.3(0.1-0.9)], while pneumothorax [OR 5.5 (1.0-29.9)], SEH[OR 11.3 (2.3, 55.9)], and BPD (defined as oxygen dependency at 28 days) [OR 10.2 (4.1-25.4)] were associated with an increased risk. The mean values for total time with PaCO2< 35 mm Hg, total time with systolic blood pressure < 35 mm Hg, and total time with pH < 7.15 were higher among cases, and mean arterial pH on the first postnatal blood gas was lower (all p < 0.05, Wilcoxon rank sum test]. When stepwise regression (with backward elimination) was used to select neonatal factors independently associated with CP, significant associations were found only for BPD [odds ratio 5.4 (2.1-14.1)] and uncomplicated SEH[odds ratio 8.3 (1.5-44.5)]. For the 27 cases of CP with normal CUS, the OR for BPD was 6.8 (2.4-19.3). If this association is assumed to be causal, then the estimated etiologic fraction for BPD is 0.66 (95% confidence limits 0.30-0.84). Thus among infants with normal CUS, a majority of CP may be attributable to BPD or unknown antecedents of BPD.