The purpose of this placebo-controlled, double-blind, randomized study was to assess the effect of nebulized morphine on dyspnoea perceived at rest by patients with advanced disease.Seventeen hospital in-patients with disabling dyspnoea received isotonic saline or morphine via nebulization for 10 min through a mouthpiece, combined with oxygen via nasal prongs. On four consecutive days, they were given one of the four following treatments in random order: saline with 2 L·min -1 oxygen; 10 mg morphine with 2 L·min -1 oxygen; 20 mg morphine with 2 L·min -1 oxygen; and 10 mg morphine without oxygen (prongs fixed, no flow). Dyspnoea was assessed on a bipolar visual analogue scale (VAS) (-100% much more short of breath, +100% much less short of breath), and arterial oxygen saturation (Sa,O 2 ) and respiratory frequency (f R) were recorded at the end of nebulization and 10 min later.In 14 subjects who completed the study, mean VAS ratings 10 min after the end of nebulization ranged +30 to +43%, with no significant difference between the four study days (VAS 20 mg morphine minus VAS saline, 95% confidence interval (95% CI) -6 to +8%). Sa,O 2 significantly increased on the 3 days with supplemental oxygen, and remained stable on the zero flow day. Respiratory frequency significantly decreased on the 4 days, with a trend to correlation between VAS rating and parallel change in respiratory frequency (Spearman's rank correlation coefficient (r s )=-0.46; p=0.09).We conclude that the subjects benefited from saline or morphine via a placebo effect and/or a nonspecific effect, and that nebulized morphine had no specific effect on dyspnoea.