Campbell, M.L., Yarandi, H., & Dove-Medows, E. (2012). Oxygen is nonbeneficial for most patients who are near death. Journal of Pain and Symptom Management, 45(3), 517-523.

The objective of the study is to determine if routine application of oxygen in patients who were near death would be beneficial as measured by relief of distress.

At baseline, a trained observer who was blinded to the current type of NC flow (oxygen or medical air) of the patient collected data about the respiratory status of each patient. Patients not in respiratory distress were then randomly alternated through low-flow oxygen, medical air, or room air on an NC on a 10-minute washout rotation. Each patient received each type of air twice for a total of six rotations per patient.

• The study reported on a sample of 32 patients who ranged in age from 56-97 years.
• The sample was 34% males and 66% females.
• Included in the study were patients who were referred for palliative care consultation, enrolled in hospice, near death, at risk for respiratory distress, and institutionalized.
• The study reported on a limited number of patients with cancer (9% lung cancer, N = 3).

The multi-site study was conducted in a university hospital or free-standing hospice facility in the Midwestern United States.

• Patients were undergoing end-of-life care.
• The study has clinical applicability for palliative care.

The double-blind study repeated measures with patient as own control.

• Respiratory distress observation scale (RDOS) - observation of respiratory distress symptoms
• Palliative performance scale (PPS) - nearness to death
• Reaction Level Scale (RLS85) to assess consciousness
• Oxygen saturation and end tidal carbon dioxide via Nellcor N-85 capnograph/oximeter (Covidien, Mansfield, MA)

Most patients (84%) had oxygen flowing at baseline. Three patients did not tolerate switching from oxygen to no flow and were restored to baseline with no further rotations. One patient died during the fourth gas/flow encounter without increased distress. No oxygen adverse effects (e.g., nosebleeds) occurred during the study. Average baseline respiratory distress was 1.47 (scale 0-4), with no difference over the study period. Baseline oxygen saturation and RDOS were inversely related. Neither consciousness nor nearness to death correlated to baseline RDOS, but consciousness was significantly correlated to nearness of death.

Routine application of oxygen does not reduce dyspnea at the end of life and should be used as an n of one trial in patients near death with observed dyspnea.

• The study had a small sample size of less than 100.
• The study had a risk of bias due to sample characteristics and no blinding.
• Key sample group differences could influence results.
• This small sample size with different medical diagnoses may not account for different types of airway/respiratory compromise and the potential variance in benefit from oxygen (e.g., heart failure, secretions, airway obstruction).

Oxygen administration near the end of life for management of observed dyspnea is scientifically logical, but this study suggests no proven clinical benefit. Costs to bring oxygen into the home and train caregivers in its administration may not be necessary based upon this study’s findings. In addition, the addition of the oxygen delivery device may be uncomfortable or produce unwanted adverse effects. The limited number of patients with cancer in this study limits generalizability to this population.