Extravascular Lung Water, B-Type Natriuretic Peptide, and Blood Volume Contraction Enable Diagnosis of Weaning-Induced Pulmonary Edema*

Abstract

We tested whether the changes in extravascular lung water indexed for ideal body weight could detect weaning-induced pulmonary edema. We also studied the diagnostic value of blood volume contraction indices and B-type natriuretic peptide variations. Prospective study. Twenty-one patients who failed a first spontaneous breathing trial. None. We performed a second 60-minute T-tube spontaneous breathing trial. Before and at the end of spontaneous breathing trial, we recorded pulmonary artery occlusion pressure, the extravascular lung water indexed for ideal body weight, plasma B-type natriuretic peptide level, hemoglobin, and plasma protein concentrations. Weaning-induced pulmonary edema was defined by the association of signs of clinical intolerance and a pulmonary artery occlusion pressure greater than or equal to 18 mm Hg at the end of spontaneous breathing trial. Because some patients performed several spontaneous breathing trial, a primary analysis included all spontaneous breathing trial and a secondary analysis included only the first spontaneous breathing trial of each patient. In primary analysis, 36 spontaneous breathing trials were analyzed, 21 spontaneous breathing trial with weaning-induced pulmonary edema and 15 without. During spontaneous breathing trial, extravascular lung water indexed for ideal body weight increased only in cases with weaning-induced pulmonary edema (25% ± 23%). Plasma protein concentration, hemoglobin concentration, and B-type natriuretic peptide also significantly increased only in cases with weaning-induced pulmonary edema (9% ± 3%, 9% ± 4%, 21% ± 23%, respectively). The areas under the receiver operating characteristics curves to detect weaning-induced pulmonary edema were 0.89 (95% CI, 0.78-0.99) for extravascular lung water indexed for ideal body weight, 0.97 (0.93-1.01) for spontaneous breathing trial-induced changes in plasma protein concentration, 0.96 (0.90-1.01) for changes in hemoglobin concentration, and 0.76 (0.60-0.93) for changes in B-type natriuretic peptide. An increase in extravascular lung water indexed for ideal body weight greater than or equal to 14% diagnosed weaning-induced pulmonary edema with a sensitivity of 67% (95% CI, 43-85%) and a specificity of 100% (95% CI, 78-100%). The secondary analysis confirmed these results. Spontaneous breathing trial-induced increases in extravascular lung water indexed for ideal body weight, plasma protein concentrations, hemoglobin concentration, and B-type natriuretic peptide are reliable alternatives to the pulmonary artery catheter for diagnosing weaning-induced pulmonary edema.