The aims of this study are to 1) determine the optimal levels of O2 flow which prevent nocturnal O2 desaturation while minimizing periods of hyperoxia during the course of nocturnal oxygen therapy (NOXT) in heart failure patients with reduced ejection fraction (HFrEF) patients with CSA/CSR; 2) document whether within-patient EO2F values change over time during NOXT, and identify factors which predict changes in EO2F; and 3) examine how well a conventional stepwise titration procedure compares to a breath by breath titration using an automated O2 titration system in terms of targeted flow rate and night time oxygenation (oxygen desaturation index, time spent at specific SpO2 targets).

Sleep-related breathing disorders (obstructive and central) are highly prevalent in Heart failure (HF) patients and are associated with an increase in morbidity and mortality. Nocturnal oxygen therapy (NOT) reduces the frequency of central breathing events by 75 % and prevents nocturnal desaturation in patients with HF. Considering that the amount of nocturnal desaturation is a better predictor of mortality than the apnea+hypopnea index (AHI) in this population, one should expect NOT to have a positive impact on survival in these patients. In the four randomized clinical trials where the effects of O2 on left ventricular function was assessed, 2 reported a significant increase in LVEF after 3 months of NOT. NOT was also found to positively impact on other important predictive factors of mortality such as sympathetic activity and VO2 max. These mitigated results could be accounted by the fact that a fixed O2 flow was empirically used (2 to 4 L/min) in the majority of studies. This may impede the beneficial effects of NOT for two reasons. First, in patients with HF, oxygen is associated with a dose-related detrimental hemodynamic effects (i.e. increase in vascular resistance and reduction in cardiac output and stroke volume). Therefore, the lowest O2 flow that prevents nocturnal desaturation should be used to minimize the detrimental effects of hyperoxia. On the other hand, there are evidences that the frequency and/or severity of sleep-disordered breathing may change overtime in CHF patients leading to insufficient correction of nocturnal desaturation during the course of NOT. Therefore, NOT should be preceded by an oxygen titration procedure to determine the lowest O2 flow that prevents nocturnal desaturation. This can be done with a stepwise night-to-night increase in O2 flow until correction of nocturnal desaturation. However, another approach would be to prevent event-by-event desaturations and to prevent hyperoxia during periods of normal sleep and wakefulness. On the other hand, the stability in O2 needs overtime in these patients is unkown. The aims of this study are 1) to document if the level of O2 flow preventing nocturnal desaturation changes during the course of NOT in CHF patients with CSA/CSR and 2) to examine the ability of automated O2 titration (FreeO2, Oxynov, Quebec, Canada) to determine O2 needs in HF patients with CSA/CSR when compared to the gold standard titration procedure.

Inclusion Criteria:

• patients with heart failure and reduced ejection fraction (LVEF < 45%) due to ischemic or hypertensive heart disease
• moderate to severe central sleep apnea/cheyne stokes respiration.
• treatment should be stable for the last 30 days preceding entry into the study.

Exclusion Criteria:

• O2 /CPAP therapy,
• active smoking,
• primary valvular heart disease,
• nasal obstruction,
• BMI ≥ 32 Kg/m2,
• cardiac surgery/transient ischemic attack/stroke/resynchronization therapy within 3 months,
• nocturnal hypoventilation,
• receiving opiates or methadone medication.