PURPOSE: To evaluate the feasibility of using chest radiography to differentiate between three different etiologies of pulmonary edema. MATERIALS AND METHODS: Plain chest radiographs of 77 patients, who were clinically confirmed as having pulmonary edema, were retrospectively reviewed. The patients were classified into three groups: group 1(cardiogenic edema : n=35), group 2(renal pulmonary edema : n=16) and group 3(permeability edema :n=26). We analyzed the radiologic findings of air bronchogram, heart size, peribronchial cuffing, septal line, pleural effusion, vascular pedicle width, pulmonary blood flow distribution and distribution of pulmonary edema. In a search for radiologic findings which would help in the differentiation of these three etiologies, each finding was assessed. RESULTS: Cardiogenic and renal pulmonary edema showed overlapping radiologic findings, except for pulmonary blood flow distribution. In cardiogenic pulmonary edema(n=35), cardiomegaly(n=29), peribronchial cuffing(n=29), inverted pulmonary blood flow distribution(n=21) and basal distribution of edema(n=20) were common. In renal pulmonary edema(n=16), cardiomegaly(n=15), balanced blood flow distribution(n=12), and central(n=9) or basal distribution of edema(n=7) were common. Permeability edema(n=26) showed different findings. Air bronchogram(n=25), normal blood flow distribution(n=14) and peripheral distribution of edema(n=21) were frequent findings, while cardiomegaly(n=7), peribronchial cuffing(n=7) and septal line(n=5) were observed in only a few cases. CONCLUSION: On plain chest radiograph, permeability edema can be differentiated from cardiogenic or renal pulmonary edema. The radiographic findings which most reliably differentiated these two etiologies were air bronchogram, distribution of pulmonary edema, peribronchial cuffing and heart size. Only blood flow distribution was useful for radiographic differentiation of cardiogenic and renal edema.