Résumé : To study adrenomedullin (AM) plasma levels in patients with severe lung disease and to analyze the relationship between AM and heart changes, hemodynamics and blood gases. Case control study of 56 patients (36 men, 20 women) with severe lung disease and 9 control subjects (7 men, 2 women). Patients with end-stage pulmonary disease, including chronic obstructive pulmonary disease (COPD, n=11), cystic fibrosis (CF, 26), idiopatic pulmonary fibrosis (ILD, n=9), and idiopatic pulmonary arterial hypertension (PAH, n=10), who were evaluated for lung trasplantation between January 1997 and September 2000, and nine patients who underwent lung surgery for a solitary benign nodule. AM plasma levels in pulmonary artery (mixed venous blood, vein) and aorta or femoral artery (arterial, art), art and vein blood gases, pulmonary hemodynamics, systemic hemodynamics, two-dimensional transthoracic echocardiography and echo-Doppler study. Plasma AM (art and ven) levels were higher among patients' group compared to the controls (AMart p<0.02 and AMven p<0.04) for CF, ILD, PAH (AMart, pg ml-1 Controls 13.7±3.6, COPD 22.8±6.2, CF 28.1±11.4, ILD 34.1±14.3, PAH 35.1±18.9; AMven, pg ml-1 Controls 14.2±4.8, COPD 28.1±12.6, CF 31.7±14.1, ILD 38.7±16.5, PAH 40.1±4.4). We found with a trend towards higher concentration in ILD and PAH patients compared to COPD and CF but no statistical significant differences. Mixed-venous AM was higher than arterial AM in all groups resulting in AM uptake (AMPulmUp pg min-1 Controls 4.8±22.6, COPD 21.1±44.9, CF 20.6±45.1, ILD 23.7±38.5, PAH 29.9±49.7). The univariate analysis showed a weak but significant correlation between AMart and mean systemic arterial pressure, heart rate, mean pulmonary arterial pressure and systemic vascular resistance. In the multivariate analysis, four variables emerged as independent factors of AMart including mean pulmonary arterial pressure, heart rate, mean systemic arterial pressure and left ventricular diastolic diameter (F=8.6, p<0.00001, r=0.60, r2=0.32). A similar weak correlation was apparent between AMven, systemic vascular resistance, and mean pulmonary arterial pressure. The results of multivariate analysis identify right atrial enlargement, mean right atrial pressure, heart rate and left ventricular dimensions as the only independent variables related to AMven (F=4.3, p<0.0004 r=0.56, r 2=0.26). AM pulmonary uptake was significantly correlated with AMven (r=0.65), but not with hemodynamic, blood gas and echocardiographic variables. AM plasma levels are elevated in patients with severe lung disease in face of a preserved pulmonary uptake. These results suggest that the high AM plasma levels in patients with severe lung disease are not caused by a reduced pulmonary clearance, instead suggesting a systemic production. © 2004 Elsevier B.V. All rights reserved.