Objective Assess the relationship between adherence to long-term oxygen therapy (LTOT) with mortality in patients with chronic obstructive pulmonary disease (COPD) and chronic respiratory failure and their clinical features. Methods Longitudinal retrospective analysis of 254 patients with COPD and chronic respiratory failure from 2008 to 2016. At baseline, we evaluated the diagnosis, spirometry values, arterial blood gas analysis, blood count, pulse oximetry, body composition and health questionnaires (dyspnea, quality of life, anxiety and depression). For referred adherence analysis to LTOT we included 199 patients, divided according to prescription of oxygen: 12h/day (G1), 15h/day (G2) and 24h/day (G3). The cause of death and dates were studied over the five-year period. Results In five years we identified 124 deaths (62.3%). No significant difference was found in mortality between the adherence groups (p=0.75) nor did we find differences in the clinical parameters evaluated. LTOT prescription was not associated with mortality (p=0.07). In Cox regression analysis, there was no association between mortality and non-adherence to LTOT (HR: 0.75; IC95%: 0.21-2.70). The risk of mortality was increased in G3 compared with G1 (HR: 7.16; IC 95%: 1.44-35.38) and in those with a higher depression score (HR: 1.35; IC: 1.14-1.59). Conclusion No association was found between LTOT adherence and mortality in patients with COPD and respiratory failure. There were no clinical differences between the adherence groups.