1. Studies in vitro have recently shown that both atrial natriuretic peptide and brain natriuretic peptide have pulmonary vasorelaxant activity. The purpose of the present study was to evaluate for the first time whether plasma levels of brain natriuretic peptide are elevated in chronic obstructive pulmonary disease. Plasma levels of brain natriuretic peptide and atrial natriuretic peptide were therefore measured in 12 patients admitted with acute hypoxaemic chronic obstructive pulmonary disease [arterial partial pressure of O2, 6.2 +/- 0.4 kPa; arterial partial pressure of CO2, 6.9 +/- 0.1 kPa; forced expiratory volume in 1 s, 0.6 +/- 0.07 litre (27 +/- 3% of predicted)]. All but three patients had oedema on admission. 2. Plasma levels of both brain natriuretic peptide and atrial natriuretic peptide were elevated in patients with chronic obstructive pulmonary disease (31.4 +/- 4.1 pmol/l and 45.0 +/- 8.1 pmol/l, respectively) compared with healthy control subjects (1.7 +/- 0.8 pmol/l and 8.0 +/- 3.5 pmol/l, respectively). Thus, plasma levels of brain natriuretic peptide and atrial natriuretic peptide in patients with chronic obstructive pulmonary disease were increased by 18.5- and 5.6-fold respectively compared with healthy control subjects. 3. There was a significant inverse correlation between the plasma level of brain natriuretic peptide and the arterial partial pressure of O2 (r = -0.65, r2 = 0.42, P = 0.03), but not between the plasma atrial natriuretic peptide level and the arterial partial pressure of O2 (r2 = 0.07, not significant). The arterial partial pressure of CO2 did not correlate with the plasma level of either brain natriuretic peptide or atrial natriuretic peptide.(ABSTRACT TRUNCATED AT 250 WORDS)