Long-term oxygen therapy (LTOT) improves survival in patients with chronic obstructive pulmonary disease (COPD) and severe hypoxaemia. However, the best method of management of moderate hypoxaemia not qualifying for LTOT (including isolated nocturnal desaturation) is uncertain. We examined the effect of home oxygen (either LTOT or nocturnal oxygen therapy) on overall survival in patients with COPD and moderate hypoxaemia. In this systematic review and meta-analysis, we searched MEDLINE, Embase, the Cochrane Central Register of Controlled Trials, CINHAL, and Web of Science from database inception to Jan 13, 2022, for parallel-group randomised trials of long-term or nocturnal oxygen in patients with COPD and moderate daytime hypoxaemia or isolated nocturnal desaturation, or both. Control groups received usual care or ambient air through sham concentrators (placebo) throughout the study period. The primary outcome of interest was 3-year mortality. Crossover trials and trials of oxygen in severe hypoxaemia were excluded. Two reviewers applied inclusion and exclusion criteria to titles and abstracts and screened the full-text articles and reference lists of relevant studies. Aggregate data were extracted manually in duplicate using structured data collection forms. Methodological quality was assessed using the Cochrane Risk of Bias tool. Random-effects meta-analysis was used to pool individual studies. We considered the minimal clinically important difference for home oxygen to be a relative risk reduction in mortality at 3-year follow-up of 30-40%. The meta-analysis is registered on PROSPERO, CRD42021225372. We identified 2192 studies and screened 1447 after removal of duplicates, of which 161 were subjected to full-text screening, and six were identified as being eligible for inclusion. These six randomised trials were published between 1992 and 2020 and the quality of evidence was high. In the primary meta-analysis (five trials; 1002 patients), we found the effect of home oxygen in reducing 3-year mortality to be small or absent (relative risk 0·91 [95% CI 0·72-1·16]; τ<sup>2</sup> = 0·00), hence the lower limit of the 95% CI did not meet the prespecified minimal clinically important difference. The results of our meta-analysis suggest that home oxygen probably makes little or no difference to 3-year mortality in patients with COPD and moderate hypoxaemia. The data do not support the widespread use of home oxygen in this patient population. None.