To quantify pulmonary contusions on chest x-ray film and to evaluate factors correlating with the size of the pulmonary contusions, changes in the first 24 hours, the need for ventilatory assistance, and death. The medical records and chest x-ray films of 103 patients with blunt chest trauma diagnosed as having a pulmonary contusion were reviewed. A pulmonary contusion score was developed (3 = one third of a lung; 9 = an entire lung). In the emergency department, pulmonary contusions were not present in 11, were mild (one ninth to two ninths of a lung) in 15 patients, moderate-severe (three ninths to nine ninths of a lung) in 53 patients, and very severe in 24 patients. Within 24 hours, the pulmonary contusion score increased in 26 patients by 7.9 +/- 5.5 (SD). The 26 patients with an increasing contusion had a higher mortality rate (38% vs. 17%) (p = 0.044) and tended to need ventilatory assistance more frequently (73% vs. 49%) (p = 0.061). The 35 patients with very severe pulmonary contusions (pulmonary contusion score = 10-18) had the lowest PaO2:FIO2 ratio at 24 hours (175 +/- 103 mm Hg), longest hospital length of stay (28 +/- 35 days), and the highest Injury Severity Score (26 +/- 9). The factors correlating highest with a need for ventilatory support (57/103) were the 24 hour or initial PaO2/FIO2 ratio < 300, an Injury Severity Score > or = 24, Revised Trauma Score < 6.4, Glasgow Coma Scale score < or = 12, and shock or need for blood in the first 24 hours (p < 0.001). Death correlated highly with a need for ventilatory assistance, Injury Severity Score > or = 26, Revised Trauma Score < or = 6.3, and Glasgow Coma Scale score < or = 11 (p < 0.001). Quantifying and noting changes in the extent of the pulmonary contusions and PaO2/FIO2 ratio during the first 24 hours may be of value in determining the need for ventilatory assistance and predicting outcome.