Appraising pulmonary edema using supine chest roentgenograms in ventilated patients

The role of portable, anteroposterior, supine chest X-rays (CXRs) in distinguishing hydrostatic pulmonary edema (HPE) from permeability pulmonary edema (PPE) in mechanically ventilated patients is controversial. We prospectively obtained and evaluated such CXRs in 33 supine, mechanically ventilated intensive-care-unit patients with pulmonary artery catheters. Three chest radiologists independently reviewed CXRs without clinical information and recorded the cardiothoracic (CT) ratio, vascular pedicle width (VPW), and other radiographic features commonly used to evaluate pulmonary edema. Hydrostatic pulmonary edema was associated with a larger CT ratio (p < 0.001), subjective impressions of cardiomegaly (p < 0.01), and increased VPW (p = 0.02). There was a significant correlation between the pulmonary artery occlusion pressure and the VPW (r = 0.45, p = 0.0076) and CT ratio (r = 0.52, p = 0.0016), as well as between the VPW and CT ratio (r = 0.49, p = 0.0032). Despite this detailed evaluation of the CXRs, the mean accuracy of the radiologists' clinical diagnosis of HPE versus PPE was 41%, and 15 of 19 (79%) of PPE patients showed one or more roentgenographic signs of volume overload. Receiver-operating-characteristic curves were constructed to determine optimum cut-off values of VPW and CT ratio associated with HPE. Hydrostatic pulmonary edema was found to correlate best using a VPW > 63 mm coupled to a CT ratio > 0.52 (p = 0.027). With this combination of objective criteria, radiologists' diagnostic accuracy could have been increased to 73%. We therefore conclude that measurements of CT ratio and VPW correlate with pulmonary artery occlusion pressure in supine, mechanically ventilated patients. Distinction of hydrostatic from permeability pulmonary edema is difficult using portable, supine CXRs, but readily assessed radiologic signs may contribute to the correct diagnosis.