Objectives: To compare the Shennan's and the consensus definition of Bronchopulmonary Dysplasia (BPD) from the National Institutes of Health (NIH) workshop and analyze specific risk factors associated with each definition. Study design: Retrospective analysis of records of 274 infants admitted to a level IV intensive care unit. Infants were classified as having BPD or no BPD by both definitions. Differences in incidence and risk factors were analyzed. Statistical methods included descriptive statistics, comparative tests, and marginal logistic regression modeling. Results: The estimated difference in prevalence was 32% [95% CI: (26%, 37%), (p < 0.0001)] between both criteria. The prevalence of BPD was 80% higher based on the NIH criteria [RR = 1.80; 95% CI: (1.58, 2.06)]. Infants with no BPD by the Shennan definition were breathing room air with or without positive or continuous pressure support and were most likely to be discharged home on oxygen [OR = 4.47, 95% CI: (1.20, 16.61), p = 0.03]. Gestational age, birth weight, and 1-min Apgar score predicted BPD by both definitions. Chorioamnionitis increased the risk of BPD by the Shennan definition but was associated with lower risk by the NIH criteria. IUGR was associated with BPD by the Shennan definition and with severe BPD by the NIH criteria. Conclusion: Compared to the Shennan's definition, the NIH consensus identified 80% more infants with BPD and is a better predictor of oxygen requirement at discharge. Until a new better criteria is develop, the NIH consensus definition should be used across centers.