In this study, we report the pulmonary edema fluid/plasma protein ratio in 10 patients in whom acute postobstructive pulmonary edema developed. The mean edema fluid/plasma protein ratio was 0.54, which is consistent with hydrostatic causes of pulmonary edema. To our knowledge, only three prior studies5,14,17 have reported the edema fluid/plasma protein ratio in postobstructive pulmonary edema. Two of the reports,5,14 also from our group, reported measurements in four patients in whom postobstructive pulmonary edema developed. These patients had edema fluid/plasma protein ratios within the hydrostatic range (mean ratio, 0.43) and were included in the current report.5,14 The other report17 had only one patient, who had a complicated hospital course, including prolonged resuscitation, cerebral infarction, and the development of ARDS. In that study, the edema fluid/plasma protein ratio was 0.83, suggesting the possibility of increased permeability pulmonary edema.17 However, the measurement may have been made some time after intubation; in the presence of intact alveolar fluid clearance mechanisms, protein concentrations in the pulmonary edema fluid will rise over time and may reach levels higher than those in plasma.9 Thus, for the classification of the etiology of pulmonary edema, it is critical to collect pulmonary edema fluid as soon as possible after endotracheal intubation. In our current study, the median time to collection of fluid after intubation was 1.5 h. The one patient in the current study (patient 10) who had an edema fluid/plasma protein ratio of > 0.75 had an unremarkable hospital course. He was extubated the same day that pulmonary edema developed and was subsequently discharged from the hospital without any prolonged pulmonary sequelae. His unremarkable hospital course makes a diagnosis of significant acute lung injury unlikely; therefore, his high edema fluid/plasma protein ratio was probably due to the absorption of lung water and the subsequent concentration of edema fluid protein prior to sampling.