Hyperoxia during outpatient pulmonary rehabilitation in chronic lung disease - does it matter?

Chronic obstructive pulmonary disease (COPD) is a global burden. Treatment options have improved in recent years, pulmonary rehabilitation plays a key role. Oxygen therapy is recommended in patients with a low saturation at rest, but no clear guidance is given for patients who desaturate during exercise. The effect of ambulatory oxygen during exercise is not yet completely understood, especially in those patients with exercise-induced desaturation. Aim: Evaluate the effect of supplemental oxygen therapy during a 3-months long ambulatory pulmonary rehabilitation program on exercise endurance. Patients will either receive supplemental oxygen therapy or room air during exercise training. Methods: A total of 32 patients with a confirmed diagnosis of COPD undergoing pulmonary rehabilitation will be included (male and female : aged ≥ 18 years; stable condition \&gt; 3 weeks (e.g. no exacerbations); resting oxygen saturation (SpO2) ≥ 88% and exercise induced hypoxemia defined by a fall in SpO2 by ≥ 4% and/ or below 90% during a 6-minute walking test). 3-months comprehensive PR for both intervention and control group including: education and awareness, instruction on inhalation therapies, smoking cessation counselling, respiratory and skeletal muscle training in groups, guided walks / cycle ergometer training. The intervention group will be trained using SSOT and the control group will receive sham air (ambient air). Patients and assessors will be blinded. Between group difference in cycling constant work rate exercise test (CWRET) time from baseline to 3 months SSOT (5l/min) vs ambient air will be the primary outcome of this study. Data will be summarized by means (SD) and medians (quartiles) for normal and non-normal distributions. Effects of treatment will be evaluated by mean differences with 95% confidence intervals, T-tests or Wilcoxon matched pair tests as appropriate. A p-value threshold of \<0.05 or a confidence interval not including zero will be considered as statistically significant. Analyses will be performed according to the intention to treat principle.