Helium has a low density and the potential of reducing the work of breathing and improving alveolar ventilation when replacing nitrogen in air. A Phase II, double-blind, randomised, prospective, controlled trial was undertaken to assess whether Heliox28 (72% He/28% O-2) compared with oxygen-enriched air (72% N-2/28% O-2) or medical air (78.9% N-2/21.19% O-2) could reduce dyspnoea and improve the exercise capability of patients with primary lung cancer and dyspnoea on exertion (Borg > 3). A total of 12 patients (seven male, five female patients, age 53-78) breathed the test gases in randomised order via a facemask and inspiratory demand valve at rest and while performing 6-min walk tests. Pulse oximetry (SaO(2)) was recorded continuously. Respiratory rate and dyspnoea ratings (Borg and VAS) were taken before and immediately post-walk. Breathing Heliox28 at rest significantly increased SaO(2) compared to oxygen-enriched air (96 +/- 2 cf. 94 +/- 2, P < 0.01). When compared to medical air, breathing Heliox28 but not oxygen-enriched air gave a significant improvement in the exercise capability (P<0.0001), SaO(2) (P<0.05) and dyspnoea scores (VAS, P<0.05) of lung cancer patients.