Randomised controlled trials that compared oxygen (delivered via nasal cannula, mouthpiece or mask) with placebo in adults with malignancy who had a mean partial arterial oxygen pressure (PaO2) of at least 55 mmHg, or at least 50% of whom had oxygen saturation ≥88% by pulse oximetry, and that reported dyspnoea outcomes, were eligible for inclusion. Studies in which patients were already receiving home oxygen therapy were excluded.
Four studies compared oxygen with medical air, and one compared Heliox28 (72% helium and 28% oxygen) with oxygen and medical air. Oxygen was delivered by nasal cannula or mask, at doses ranging from 3 to 5 litres/minute. Oxygen was administered at rest or during a 6-minute walk test.
The included patients had a median age of 65 years and 39% were female. The patients had lung cancer, unspecified cancer with metastasis to the lung, breast cancer, colon cancer, sarcoma, carcinoid cancer, skin cancer, bladder cancer, and head and neck cancers. Baseline oxygen saturation, where reported, ranged from a median of 93 to 98%; one study reported a range of 80 to 99%.
Dyspnoea was assessed using 100-mm visual analogue scales alone or combined with modified Borg ratings, or using a 0 to 10 numerical rating scale.
Two reviewers assessed studies for inclusion, and any disagreements were resolved through consensus.