During the last decade topical glucocorticoids have been established as the first hand choice for the treatment of allergic rhinitis. Although they are clinically effective, their precise mode of action has not been sufficiently clarified. In order to evaluate whether the clinical effect is dependent on the topical administration a randomized, double-blind, double-dummy study was performed. Ninety-eight patients with allergic rhinitis due to birch pollen were recruited at two centres in southern Sweden. The patients received one of the following three treatment alternatives: 200 μg of nasally applied budesonide twice daily, 250 μg of budesonide given orally twice daily or a placebo. The selection of the doses was based on previous pharmacokinetic studies giving almost equal plasma levels of the nasally and orally administered budesonide. To evaluate the possible clinical efficacy of the treatment, the patients kept a diary in order to register symptoms according to a 0–4 scale. After a run-in week with no treatment the patients entered the treatment period which lasted for 3 weeks. A total of 96 patients completed the study. The symptom data showed that topically applied budesonide was clinically effective with a pronounced reduction in all the nasal symptoms registered and compared with placebo treatment there was a statistically significant difference in favour of the budesonide treatment (P < 0.001). The nasal symptoms of the patients receiving oral budesonide did not differ from those receiving placebo. Intranasally administered budesonide proved to be significantly more effective than orally applied budesonide (P < 0.01). Thus, this study clearly demonstrates that the clinical effect of budesonide in recommended doses for hay fever treatment is dependent on the nasal application.