Many patients with chronic obstructive pulmonary disease (COPD) receiving supplemental oxygen state that this treatment makes them less short of breath at rest. We postulated that this phenomenon may be related to improved arterial oxygenation, reduced ventilation, or stimulation of nasal receptors caused by the flow of gas. Eight patients who reported this phenomenon were studied in a quiet room. Each patient received zero flow, 2, or 4 L/min of air or oxygen through nasal cannula for 5 min at each level in random order in a single blind manner. At the end of each period, arterial blood gas composition was measured, and breathlessness was assessed with a visual analog scale. The scale was calibrated to read from zero (not at all breathless) to 100 (extremely short of breath). The entire protocol was repeated after application of topical lidocaine to the nasal passages. Results were assessed by analysis of variance. We found no significant effect of inspired oxygen concentration, gas flow, arterial oxygen tension, or arterial carbon dioxide tension on breathlessness. There was, however, a significant increase in breathlessness after nasal anesthesia from 44 +/- 3 SEM to 52 +/- 4 SEM (p less than 0.005). We suggest that the reduction of breathlessness in these patients by nasal oxygen is a placebo effect caused by wearing the nasal cannulas and is unrelated to gas flow or the increased arterial oxygen tension.