To determine the rate of magnetic resonance imaging (MRI)-detected noncystic white matter injury (WMI) in a prospective cohort of premature newborns, and to evaluate its associations with changes in clinical predictors of WMI over the study period.A prospective cohort of premature newborns (<33 weeks gestational age) was studied with MRI within 4 weeks of birth and near term-equivalent age. A pediatric neuroradiologist scored the severity of WMI on T1-weighted MRI according to published criteria. WMI was classified as none/mild or moderate/severe. Subjects with severe cystic WMI, periventricular hemorrhagic infarction, or motion artifact on MRI were excluded. Changes in clinical characteristics and predictors of WMI over the study period (1998-2011) were evaluated. Predictors of moderate/severe WMI, including birth year, were evaluated using multivariate logistic regression.Among 267 newborns, 45 (17%) had moderate/severe WMI. The rate of moderate/severe WMI decreased over the study period (P = .002, χ(2) test for trends). On multivariate logistic regression, the odds of moderate/severe WMI decreased by 11% for each birth year of the cohort (OR, 0.89; 95% CI, 0.81-0.98; P = .02). Prolonged exposure to indomethacin also was independently associated with reduced odds of moderate/severe WMI.The decreasing burden of MRI-detected moderate/severe noncystic WMI in our cohort of premature newborns is independent over time of changes in the known clinical predictors of WMI. Prolonged exposure to indomethacin is associated with reduced WMI.