Cheyne-Stokes respiration (CSR) during sleep is common in patients with severe congestive heart failure (CHF). It induces repetitive oxygen desaturations and impairs sleep [1,2]. Disturbed sleep is likely to cause daytime symptoms and the repetitive oxygen desaturations and arousals increase sympathetic activity as well as right and left ventricular afterload [3] and may thus further impede left ventricular function and exercise tolerance [4][5][6]. Effective treatment for CSR is therefore needed. Nocturnal oxygen by nasal prongs reduces CSR by about 50% and consolidates sleep [1,[6][7][8]. Application of 3% CO 2 prevented CSR by increasing the arterial carbon dioxide tension (Pa,CO 2 ) above the apnoeic threshold [9] but sleep was adversely affected, this being attributed to the tight-fitting face mask used [10]. The hypothesis was tested that CO 2 in conjunction with O 2 given by nasal prongs is efficacious in the treatment of CSR. Plasma catecholamines were measured, to evaluate possible effects on sympathetic activity. Methods Subjects and protocolAll patients with severe heart failure admitted to the department of cardiology were candidates for the study.Patients under the age of 75 yrs were eligible if they met the following criteria: at least one episode of cardiac decompensation, ejection fraction ð35%, stable condition on cardiac medication and evidence of CSR by nocturnal polysomnography. Exclusion criteria were myocardial infarction within 1 yr of entry, significant obstructive lung disease as defined by a forced expiratory volume in one second (FEV1)/forced vital capacity (FVC) <70%, primary valvular heart disease, tibial oedema or evidence of obstructive sleep apnoea (more than five obstructive apnoeas·h -1 ). Plasma catecholamines were evaluated in 15 healthy subjects (age 53.2±5.3 yrs) without significant sleep-disordered breathing. The study was approved by the local ethics committee. Informed written consent was obtained from all subjects.The study was designed as a single-blind, placebo-controlled trial. After an accommodation night where no treatment was applied the patients received air as well as O 2 plus CO 2 in a randomized, cross-over fashion on two consecutive nights. Polysomnography was performed during all three nights. Oxygen was given with a flow rate of 2 L·min -1 by nasal prongs and CO 2 was admixed simultaneously with a flow rate of 0.2-1 L·min -1 . Accordingly, the flow rate of the mixture was 2.2-3 L·min -1 . CO 2 flow rate was regulated depending on the transcutaneous carbon dioxide tension (Ptc,CO 2 ), which was not allowed to increase >7.3 kPa (55 mmHg) and on the occurrence of In conclusion, nocturnal O 2 plus CO 2 improves Cheyne-Stokes respiration in patients with congestive heart failure but has adverse effects on the sequel of CheyneStokes respiration, namely sympathetic activation. Eur Respir J 1998; 12: 414-419.