We compared the clinical effect of the glucocorticoid budesonide delivered from two nebulizers Aiolos and Pari LL in 38 children less than 4 years of age (mean age, 20.2 months) with chronic wheeze. The design was a controlled, single-blind, randomized, cross-over, dose titration study. After a 1-week run-in, patients were randomized to treatment with 1 mg budesonide b.i.d. for 2 weeks from either an Aiolos or a Pari LL nebulizer. This was followed by a gradual dose reduction period during which the budesonide dose was reduced at 2-week intervals until unacceptable asthma symptoms appeared or the placebo level was reached. The patient was then switched to budesonide 1 mg b.i.d. from the other nebulizer for 2 weeks, after which the dose was reduced at 2-week intervals as described for the first period. Patients who completed the study on placebo for 2 weeks without deterioration of their asthma were not included in the statistical analysis. During Period #1 the minimum effective dose of budesonide was 2 mg/day in 9 patients, 1 mg/day in 10 patients, and 0.5 mg/day in 13 patients. In Period #2 the corresponding figures were 14, 5, and 13 patients. Six patients were excluded after the first period because their asthma control did not deteriorate during dose reduction and when finishing on placebo for 2 weeks. For both nebulizers the reduction in budesonide dose was associated with a small increase in symptoms and use of rescue terbutaline. The mean dose of budesonide delivered to the patient by the Aiolos was twice as large as that delivered by the Pari LL: 26% vs. 13% of the nominal dose assessed by the filter method. Nevertheless, no statistically significant difference in clinical effect or mean minimal effective dose (1.1 mg for Aiolos and 1.2 mg for Pari LL) could be detected between the two nebulizers. No serious adverse events were observed. We conclude that the minimal effective dose of nebulized budesonide varies from 0.5 to 2.0 mg/day in young children with asthma. A higher drug delivery, as assessed by the filter method, does not necessarily result in better clinical control or lower minimal effective dose. Further studies are needed to assess whether this is due to insufficient sensitivity of the study design in detecting a difference in clinical effect, or whether measurements of drug delivery by the filter method do not reflect lung deposition or clinical effect in young children with wheezing.