A retrospective analysis (1980-1990) of normally formed low birthweight (< 2500 g) infants surviving to at least 28 days following intermittent positive pressure ventilation (IPPV) for longer than 12 h was performed. Bronchopulmonary dysplasia (BPD) was defined as oxygen dependency at 28 days with characteristic radiographic findings. Logistic regression analysis of risk factors, before and after the initiation of IPPV was performed on 412 infants. Decreasing birth weight (BW) and gestational age (GA) were associated with an increased risk of BPD. When controlled for these variables, predictive factors prior to IPPV were gender, age at IPPV, respiratory diagnosis, and year of birth. Following IPPV, duration of peak inspiratory pressure > 25 cm H2O, duration of fraction of inspired oxygen (FiO2) > 0.60 (DO2), maximum peak inspiratory pressure (MPIP), maximum FiO2, patent ductus arteriosus, bacteraemia and either pneumothorax or pulmonary interstitial emphysema were associated with an increased risk of BPD. Adjusting for BW and GA, there was a significant reduction in BPD risk from 1980-1990 (relative odds of 0.88 for each year compared to the previous year). This trend could be largely accounted for by decreases in MPIP and DO2 during the study period. Surfactant treatment was not independently associated with a significant change in the risk of BPD. Based on this analysis, we developed a scoring system for predicting the risk of BPD in the neonatal period which we evaluated in a random sample of infants. This predicted infants at risk of BPD with a sensitivity of 65% and a specificity of 88%.(ABSTRACT TRUNCATED AT 250 WORDS)